A top Microsoft exec has admitted that Windows 11 still relies on a bunch of old code from the 1990s. It is refreshing for Microsoft Azure Chief Technical Officer Mark Russinovich to highlight this fact on social media, but it might not surprise as many folks as he thinks.

As Russinovich eloquently puts it, those of us invested in the computer scene in the 90s “were thinking flying cars and moon stations by the year 2026, not Win32.” The admission that such old software tech is still the "bedrock" of Windows today may be the CTO strategically sharing a cold, hard truth, providing a 'let's be real' moment as part of Microsoft’s latest charm offensive. Sharing a candid confession indicates that the corporation is actually aware of the issues in its OS.

Remember, the firm is currently in the midst of a major transformation, targeting enthusiast hot button areas like Windows performance, overhead, and reliability. This drastic pivot was cautiously welcomed in contrast to Microsoft being widely slammed for boasting about Windows “evolving into an agentic OS” last November. Currently, Microsoft seems to be flailing around, trying to stop folks straying to pastures greener like Mac and Linux.

However, Win32 isn’t inherently labeled as ‘bad’ by the Microsoft CTO. Though it is ancient, it has probably stuck around for good reason. “I think one of the reasons it’s got this staying power is just a fundamental layer inside of Windows that so many apps have built on,” notes Russinovich. “So many technologies and ecosystems have been built on top of it that it’s kind of a bedrock.”

The CTO explains that Win32 has persisted even when facing targeted existential threats from within Microsoft, particularly in the Windows 8 era. “There’s been various times in Microsoft’s history where we thought we’d reboot the Windows API surface, like WinRT, that actually didn’t play out the way a lot of people expected it to.”

In closing, Russinovich highlights that Win32 was also the bedrock for tools like Sysmon and ZoomIt, which he actually wrote back in 1996. These tools are now “more relevant than ever in 2026,” as parts of Windows 11 and PowerToys, respectively, reckons the CTO.

A performance-crushing bug in Monster Hunter Wilds appears to vanish once all of the game's DLC is purchased and added to the game. Redditor de_Tylmarande highlighted this strange issue with the game, where DLC-rich folks can enjoy double the FPS compared to base game owners, and has been in touch with Capcom developer contacts.

Coincidentally, Capcom has just announced a performance patch is coming to the game, designed to reduce PC Steam “processing load.” The OP’s investigations pointed to aggressive background DLC ownership checks being the central performance issue.

de_Tylmarande was unhappy with Monster Hunter Wilds performance, explaining that “constant FPS drops where they should not exist at all were driving me crazy.” The OP had previously managed to iron out performance bugs in Capcom’s Dragon's Dogma 2, in collaboration with one of the company devs, so felt inspired to try again.

Initial investigations pointed to some sporadic CPU-sapping background process. In the OP’s case, it meant that the significant FPS dips were fixable by switching the gaming laptop to performance mode. But they weren’t satisfied with this noisy solution.

Research into the issue continued, but a breakthrough came by “pure accident.” de_Tylmarande was running the game using a friend’s Steam account (on the same laptop) and “everything was great.” They then spent some time making sure settings like DLSS, HD textures, caches, and other things remained the same when switching between accounts.

As the only difference began to become clear, “I refused to believe what I was seeing,” says the OP. The inescapable conclusion of the investigation was, “the more DLC you own, the better performance you get in the game.” Sherlock would be proud.

The performance delta isn’t negligible. de_Tylmarande experienced half/ double the performance between accounts with the base and full-DLC-loaded games. In A-B testing, “on the account with no DLC I get heavy and stable FPS drops down to like 20-25 in hubs, while on the account with all DLC bought it's 80+ FPS,” it is claimed.

Proving up the +DLC for +FPS theory

Carrying their investigations one further step forward, and so sniffing out a distributable solution in case Capcom doesn’t fix the issue, the Redditor created a mod to emulate owning all the DLC. de_Tylmarande stresses that this mod isn’t a hack to unlock unpaid for DLC.

With the mod loaded, “performance went through the roof.” Fooling the game into thinking you own all the DLC was a winning performance tuning strategy. The reason, reckons the OP, is simple. Performance was confirmed by the mod to be sapped by the “insanely crooked and aggressive DLC ownership check function” that Capcom has in place

After talking to Capcom contacts about their findings, de_Tylmarande says, “I hope we will see a huge performance fix soon.” The Redditor is not suggesting there was any bad intent on Capcom’s side, but just wants to highlight this bug and get it fixed for all users.

Indeed, we see the official Monster Hunter Status social media accounts today (embedded top) confirm that “Patch Ver.1.040.03.01 is set to go live on Steam January 27, 6 pm PST / January 28, 2 am GMT.” It is not specified that it addresses this particular wrinkle, but the wording “This patch will include optimization improvements for Steam-specific processes and options to reduce processing load,” certainly sounds like it.

There will be more stability and performance updates across all platforms, from a subsequent Ver. 1.041 update, released on February 18.

We have reached out to Capcom to get a definitive statement about the DLC check performance issues.

Windows 11 gets a bad rep in the community because of its higher baseline overhead, stringent hardware requirements, UI regressions, and more - not to mention the forced Microsoft hooks that keep getting worse by the day. Moreover, when placed in a rather unscientific test by TrigrZolt, comparing six different generations of Windows with each other, it placed dead last in pretty much every individual test, though the situation is a bit more nuanced.

Six Lenovo ThinkPad X220 laptops were used in the test, featuring a Core i5-2520M CPU and 8GB of RAM, with a 256GB hard drive — running the latest versions of Windows XP, Windows Vista, Windows 7, Windows 8.1, Windows 10, and Windows 11. That setup alone should tell you how the methodology employed here is skewed toward favoring older software. Windows 11 isn't even officially supported on these components.

Regardless, the experiment begins with a startup test, where Windows 8.1 booted up the quickest, while Windows 11 was the slowest. Both of these versions (and Windows 10) have Fast Boot capability that older Windows editions lack. In the video, we also see that Windows 11 struggles to load the taskbar for a bit, an infamous quirk of the OS that's been heavily scrutinized ever since launch.

Then comes the storage test, where it's Windows XP that takes the cake, with only 18.9GB of space taken up for all the apps installed. The same number of programs, along with Windows itself, took 37.3GB of hard drive real estate on Windows 11, so there's definitely a lot of extras there. But Windows 11 actually came third here, behind Windows Vista, at 37.8GB, and the revered Windows 7, at a whopping 44.6GB.

Next up is RAM management where Windows XP is the winner once again, consuming only 0.8GB of system memory at idle, while Windows 11's appetite grew to 3.3GB on average; it jumped to 3.7GB at one point. This is because of the added resources the OS loads in the background, including persistent telemetry.

Older hardware with less RAM, therefore, will be more susceptible to sluggishness on Windows 11. Keep in mind, TrigrZolt is also running a system with a hard drive, which are outdated at this point regardless of your operating system loyalties. Any modern system with a decent CPU and NVMe SSD will likely mask over the general inefficiency Windows 11 shows, plus options like debloat tools and Xbox FSE can further help here.

Now we move on to the second part of the RAM management test, where the YouTuber loaded as many browser tabs as possible before the memory hits 5GB of utilization. Since Firefox and Chrome don't load webpages properly anymore on archaic Windows versions, a more widely-compatible browser called Supermium was used across all devices.

Once again, Windows 11 places dead last here, only being able to load a measly 49 tabs. Compare that to the insane 252 tabs Windows 8.1 was able to load. Even the older Windows XP managed 50 tabs, and that's because it kept crashing past that number because of its paging file failing to keep up, not because it had hit the 5GB memory ceiling.

Our fourth test is for battery life and, of course, Windows 11 died first here, while Windows XP walked away with the best endurance. Though, the delta between all the devices was only about two minutes so it won't make a difference in real-world usage. All the laptops had 100% battery health, too, and the same program was run to drain them as quickly as possible.

Moving on, exporting an audio file in Audacity once again put Windows 11 at fifth place, only ahead of Windows Vista which was experiencing an unusual delay, otherwise all laptops finished around the same time. The same fate follows Windows 11 when it came to rendering a video, finishing in last, with Windows 10 taking first place. Here, Windows XP and Vista couldn't load the OpenShot Video Editor that was used, so they were disqualified.

In application opening times, Windows 11 got last place across all five programs that were tested: File Explorer, MS Paint, Calculator, Adobe Reader, and VLC Media Player. Older versions of Adobe Reader and VLC were used to ensure compatibility with all six operating systems, so there's a bit of performance left on the table, but still the native apps didn't win any awards either.

After so many consecutive losses, Windows 11 actually secured third place in one half of the web browsing test where it had to load an image, but fell to last place again when visiting the Google Images and Microsoft Account login websites. When transferring files, though, Windows 11 snatched second place, only behind Windows 10, while placing fourth in the malware scan test using MalwareBytes — Windows 7 won this one.

Finally, we arrive at our last test, which are benchmarks. Windows XP took the crown in CPU-Z's single threaded load while Windows 7 topped the multithreaded charts, and Windows 11 was fourth in both. Geekbench was also tested, but the versions are different for Windows 10/11 and the rest. Between those two, Windows 11 scored higher in single-core but lower in multicore. Among the older operating systems, Windows Vista walked away with the highest score.

TrigrZolt also tested CrystalDiskMark and Windows 11 netted a decent third place here, tied with Windows 10, while Windows XP won. In Cinebench R10 single-core, Windows 8.1 got its second victory of the day, pushing Windows 11 down to fourth place. Multicore was even worse for Windows 11 because it only beat Windows 10 by a few points to save itself from scoring last; Windows Vista walked away as the fastest.

All in all, this was a pretty devastating showing for Windows 11 where it couldn't even win a single test. However, the hardware is just so outdated at this point that it doesn't represent the Windows 11 experience faithfully. The laptops were never designed to run a modern operating system, neither does Microsoft's latest OS support this hardware.

If you were to use modern laptops that came out in the Windows 10/11 era, the results would likely be skewed the other way. The omission of an SSD, especially, is strange since that's a component every edition of Windows will benefit from, and something that truly does hold back Windows 11 here. Though, it's still a little embarrassing, considering Windows 10 didn't suffer nearly as much without it.

The YouTuber himself mentions that this test represents historic value more than practicality. He ended up giving the overall win to Windows 8.1, saying how fluid and fast it felt, despite being ridiculed at the time of its release. It's also more visually consistent with Windows 10/11 than it is with Windows 7/Vista, so it doesn't look outdated.

A better methodology would've been to use flagship — or even midrange, for that matter — laptops from every generation: custodians of that era of Windows, so that each version had the best shot at performing at its full potential. As it stands right now, while it's quite funny to see Microsoft's increasingly AI-riddled OS loose against legacy offerings, the test just wasn't set up fairly to begin with.

SpaceX revealed its new V3 Starlink satellites, which it says will provide gigabit connectivity to users and enable 60 Tera-bits-per-second downlink capacity on the Starlink network. The company says each launch will add 20 times the capacity of its current V2 mini offering.

The above animation was first shared by SpaceX on Monday, during a live stream of the 11^th flight test of SpaceX’s reusable Starship vehicle, notes PCMag. And now we can see why SpaceX has started to use its more powerful Starship vehicle, in preference over the existing Falcon 9 rocket, which was used for prior generation satellite deployments..

As well as their obviously significantly larger physical form, a V3 satellite is estimated to weigh in at as much as 2,000kg (4,409 pounds). Compare that to the existing V2 Mini satellites, which are less than 600kg, and the V1 satellites at around 300kg.

We’ve established that these V3 Starlink satellites are big and heavy, so what are the network performance gains? SpaceX’s latest social media bulletin regarding the performance of these units is quite brief. However, it highlights that V3 will “bring gigabit connectivity to users and are designed to add 60 Tera-bits-per-second of downlink capacity to the Starlink network.”

Thankfully, there were some prior news releases and filings, which flesh out V3 details. An important one, for example, is that for every upcoming Starship launch, it will be possible to put 60 V3 satellites into orbit. “That's more than 20 times the capacity added with every V2 Mini launch on Falcon 9,” says SpaceX. Moreover, each new V3 satellite boasts 1,000 Gbps of download and 200 Gbps of upload bandwidth, making them better than 10x faster than V2 models.

With the V3 satellites deploying, probably early next year, we might begin to see SpaceX CEO Elon Musk’s original vision of a global broadband service materialize. The new satellites’ next-gen architecture should deliver the increased coverage and speed to iron out connectivity and capacity wrinkles some users are experiencing with the service, as it stands. However, it has previously been indicated that Starlink users will need new hardware to benefit from the best speeds V3 satellites will enable.

Hopefully, SpaceX is still designing even its latest, biggest, and heaviest satellites to completely burn up upon reentry, at the end of their working lives. With V3, there’s a lot more mass to be combusted on the way down to Earth, and as more satellites head into orbit, space debris — and that debris falling back to earth — remains a growing concern.

A computer scientist used only “pure SQL” to construct a multiplayer DOOM-like game. The resulting first-person shooter game, cobbled from a mere ~150 lines of Python code, is dubbed DOOMQL. Despite the self-imposed software architecture restrictions, Lukas Vogel, co-founder of database performance outfit CedarDB, says DOOMQL plays at “a breezy ~30 FPS.” It isn’t the most graphically splendid DOOM-inspired game, though.

Vogel’s SQL-powered DOOM dreams were precipitated by Patrick Trainer’s DuckDB-DOOM. This earlier experiment in code also sought to create a DOOM-like first-person shooter in SQL. However, Trainer's code, which runs in a single web page in a browser using WebAssembly, also uses JavaScript for rendering and inputs.

Other key differences in DOOMQL and the DuckDB-DOOM clone are that the former is a multiplayer effort, with SQL used for both rendering and input, achieving ~30 FPS at 128 x 64 pixels resolution. Vogel says, without a hint of scorn regarding Trainer’s efforts, that “having parts of the rendering pipeline in JavaScript felt like cheating.” Despite the ‘cheating,’ the DuckDB-DOOM clone only runs at around 8 FPS, and it uses a lower resolution. Making Vogel’s feat all the more impressive is the claim that he coded DOOMQL during a single month while on parental leave.

You can deep dive into Vogel’s SQL-powered DOOM via the linked blog post, and check out the code on GitHub. For a brief overview of the game architecture, though, it can be surmised at the high-level by the following quartet of bullet-points:

• State lives in tables (map, players, mobs, inputs, configs, sprites, …)
• Rendering is a stack of SQL views that implement raycasting and sprite projection.
• The game loop is a tiny shell script that executes a SQL file ~ 30 times per second.
• The client is ~ 150 lines of Python: It polls for input and queries the database for your 3D view.

Tantalizing Battlefield 6 morsels have been making the rounds ever since the title's official announcement last month. The iconic shooter franchise is poised to make its return this year with a back-to-basics approach, something sorely missing from contemporary FPS titles. Add to that the disappointing response to COD: Black Ops 6 and EA has found themselves stirring the perfect concoction of demand for an AAA game launch. A couple of days ago, it revealed the relatively modest hardware requirements for the upcoming Battlefield 6 and, today, we allegedly have our first performance report to pick over.

Popular Twitch steamer, Bruhskey, took to social media to show off the game's average performance. Running at 1440p, they were able to get over 300 FPS on a setup consisting of an AMD Ryzen 9800X3D CPU and an RTX 5080 GPU — both attractive top-end components. Upon this foundation was 32GB of 6400MT/s DDR5 memory running in a 1:1 ratio, which explains the excellent 6.7-8.5 total system latency the streamer was able to achieve. The game is still under embargo so they couldn't say much, like whether any upscaling was used to hit those numbers, but in general ~300 FPS in a modern blockbuster title remains impressive.

Third-person action-adventure title MindsEye had a disastrous launch after many day-one players complained of performance issues despite driving top-of-the-line equipment. One creator shared on X (formerly Twitter) that they have an average frames per second (fps) of just 55, despite running an RTX 5090 paired with an AMD Ryzen 7 9800X3D and 64 GB of DDR5. The screenshot shows the game's 98% GPU usage and gobbles up 19 GB of VRAM. It comes after the game was universally panned on launch for its unfinished state, with players reporting wild glitches and strange NPC behavior.

One player running a more modest system—an RX 9070 XT GPU matched to a 9800X3D processor—tried the game set to the lowest possible quality. They were able to achieve a more respectable 84 fps, but they had atrocious 21 fps 1% lows and 16 fps 0.1% lows.

These absolutely horrendous numbers make the game quite unplayable on most gaming PCs, and it’s quite reminiscent of Cyberpunk 2077, which ran into the same problems when it launched five years ago. One might argue that the developers would’ve warned the players about the hardware requirements for MindsEye, but the Steam page says otherwise.

The game’s minimum requirements are an Intel Core i5-12400F or an AMD Ryzen 5 5600X matched to a GPU with 6 GB of VRAM, like RTX 2060 or RX 5600 XT, and 16 GB of RAM. If you want a better experience, its recommended specifications are an Intel Core i7-13700K or AMD Ryzen 7 7800X3D with a GPU like the RTX 4070 or RX 6800 XT. However, the samples we’ve seen show that these specifications are largely unable to run the game at good quality with acceptable fps numbers.

Phison has been showcasing its data storage innovations at Computex 2025. Of particular interest, we attended a partner demo where the firm utilized 32 of its latest and greatest PCIe Gen5 SSDs in a RAID setup. During the demo, we saw CrystalDiskMark report data transfer rates as high as 113.6 GB/s for reads, and 104.6 GB/s for writes.

The PS5028 for AI Workstation Live Demo

This demonstration of extraordinary speed was enabled by a powerful modern workstation PC and a mix of cutting-edge storage components. Specifically, Phison utilized an AMD Ryzen Threadripper Pro 7978WX CPU, installed on an Asus Pro WS WRX90E-Sage SE motherboard, as the foundation of the system.

The storage subsystem consisted of 32 of Phison’s potent new E28 Gen5 SSDs installed across a trio of Apex Storage X16 Gen5 add-in cards. All these components were installed in a sweet-looking be quiet! chassis, as you can see.

Phison’s new E28 Gen5 SSD rubs shoulders with the best in consumer mid-2025 PC land. A single E28 Gen5 in a modern PC will be able to deliver data transfer speeds of up to 14.8 GB/s reads and 14.0 GB/s writes. The Phison E28 Gen5 is being marketed as “the weapon of choice for serious gaming and productivity,” so it is definitely targeting consumers. Its controller features a quad-CPU architecture, is fabbed on TSMC’s 6nm node, and supports up to 32 TB.

Here we had 32 of Phison’s new E28 Gen5 SSDs; however, each Apex Storage X16 Gen5 add-in card can fit 16 M.2 SSDs (there’s a clue in the name), so two would have sufficed for a 32x SSD demo. We guess three cards were installed for optimal performance, load balancing, or another nuance of the Threadripper-powered Windows system.

The storage performance in this setup seems remarkable. Witnessed data transfer rates as high as 113.6 GB/s for reads, and 104.6 GB/s for writes might seem amazing. However, it was confirmed by a Phison rep that the Windows kernel was actually holding back performance.

AMD reportedly held a press briefing and disclosed more information about its upcoming Radeon RX 9000-series graphics processors as well as the RDNA 4 architecture. Perhaps the most important part was the disclosure of AMD's official performance numbers of the new Radeon RX 9070 XT graphics card that appeared to be significantly ahead of the Radeon RX 7900 GRE, according to the allegedly official numbers published by VideoCardz.

In fact, AMD claims that the upcoming Radeon RX 9070 XT is 42% – 168% faster than the Radeon RX 7900 GRE at a 4K resolution with 'ultra' quality settings across over 30 games. The Radeon RX 9070 XT outperforms the RX 7900 GRE by an average of 38% at 1440p and 42% at 2160p. However with certain titles that rely on ray tracing more than others — such as Cyberpunk 2077 and Hitman 3 — performance gains reach 164% –168%, again according to the numbers published by VideoCardz.

Games with ray tracing tend to see the biggest increases, emphasizing AMD's RDNA 4 advances in handling RT workloads. Titles like Cyberpunk 2077, Dying Light 2, F1 24, and Hitman 3 show the strongest performance jumps of 56% to 66%, which clearly makes the new Radeon RX 9070-series offerings strong contenders to sit amongst the best graphics cards in the coming quarters.

When it comes to the performance difference between the Radeon RX 9070 XT and the Radeon RX 9070 (non-XT), the delta averages between 16.1% at 1440p Ultra Settings and 18.3% at 2160p Ultra Settings across the tested games, according to AMD. Yet, the Radeon RX 9070 (non-XT) still delivers a 20% performance boost at 1440p and a 21% higher performance at 2160p over the Radeon RX 7900 GRE.

AMD admitted that it did not have a GeForce RTX 5070 Ti for comparison, so it did not compare its new flagship against some of Nvidia's most wanted parts of today. The company did not explain why it decided not to compare its upcoming Radeon RX 9070 XT against its existing Radeon RX 7900 XTX flagship, but stuck to the cut-down Radeon RX 7900 GRE. The latter has around 25% lower compute performance compared to the range-topping Radeon RX 7900 XTX, but also has 16GB of memory onboard, whereas the range-topping board carries 24GB of GDDR6 VRAM.

Unlike AMD, Nvidia uses upscaling technologies like frame generation to demonstrate performance improvements over the previous generation, so the red company gains some kudos. As a result, AMD has chosen to focus on native rendering performance and ray tracing, so performance gains are quite real. More details will, of course, be shared on February 28 when AMD officially presents its Radeon RX 9070-series products.

Introducing the Nvidia GeForce RTX 5070 Ti

The Nvidia GeForce RTX 5070 Ti marks the third entry for the Blackwell RTX 50-series GPUs, which Nvidia officially unveiled at CES 2025 in early January. It takes over from both the original RTX 4070 Ti as well as the newer RTX 4070 Ti Super, with the latter having replaced the 4070 Ti last January. The 5070 Ti looks better against the 4070 Ti, while it's more of an incremental upgrade from the 4070 Ti Super. The good news is that its MSRP is also $50 cheaper than the launch prices on the 4070 Ti / Super cards, coming in at $749. Faster, cheaper, and new features is a great way to make it onto our list of the best graphics cards — assuming there's sufficient supply, which remains to be seen.

Nvidia is allowing reviews of the base-MSRP 5070 Ti models today, with higher-priced variants tomorrow (and after). The official launch date for all RTX 5070 Ti cards is February 20, 2025, as well, so you can't buy one until tomorrow. We were told that there should be quite a lot more RTX 5070 Ti cards at launch than either the RTX 5090 or RTX 5080, but even so, we anticipate the first batch at least will sell out quickly, and prices will likely head north in the short term before coming back down (we hope).

You might think that a month after the RTX 5090 launch, things would have quieted down, but that's not really true. In addition to trying to test some third-party AIB (add-in board) cards, we've had several new drivers, the 5080 launches, and unfortunately, one of our sample AIB cards has been finicky (to say the least). We're still looking for a solution, which may involve sending the card back for a different sample. The 50-series launch hasn't gone smoothly, in other words.

We did a lengthier deep-dive into DLSS 4 and MFG, using the 5080 and 5090, to get a better understanding of what the tech does and doesn't offer. The short summary is that MFG is a lot like framegen, with even more marketing hype. When it gives a straight doubling in performance, from 60 to 120 fps, or from 120 to 240 fps in the case of MFG, you can make a good argument that it looks and feels "better." But typically, it's more like a 50~80 percent improvement, making comparisons more difficult, and if it's less than a 50% increase, it can end up feeling worse.

We haven't had sufficient time to do additional MFG testing on the 5070 Ti (yet), but we'll flesh out that section of this review in the coming days. While the official Nvidia party line is that MFG makes the 5070 Ti "twice as fast" as the previous generation, that's a gross exaggeration. It might be able to spit out twice as many frames — 1x rendered and 3x generated — compared to the 4070 Ti, but the actual feel of games with MFG doesn't improve nearly as much as those inflated numbers would suggest.

For additional information about Nvidia's latest Blackwell GPUs, check the links in the boxout. The RTX 5070 Ti continues the same pattern, just with fewer GPU cores and less performance than the 5080 and 5090. It offers the same feature set as the other Blackwell GPUs, like FP4 number format support (for AI) and MFG (Multi Frame Generation) for gaming. It also has the same 16GB of GDDR7 memory as the 5080, though the memory is clocked slightly lower. Let's start with the specs table and discuss how it stacks up.

In terms of raw specs, the big changes are in memory speed (and capacity versus the 4070 Ti) plus the FP4 tensor support. There are other architectural differences that we've discussed elsewhere, like how the CUDA cores are now all "full citizens" and support both FP32 and INT32 (only half the cores in Ada and Ampere supported INT32), and there are other changes to support new neural rendering techniques. But forget all that for a moment and just look at the die sizes and transistor counts.

RTX 5070 Ti offers a decent step up from the original 4070 Ti, but compared to the 4070 Ti Super, it looks extremely similar. They both have the same number of transistors and die size, using the same TSMC 4N process node. The transistor counts aren't necessarily 100% accurate, as they're more of a rough estimate, but it does suggest that the architectural changes may not be that significant. It's also possible some things were dropped or simplified to make room for new features. What those things might be is more difficult to pin down.

Getting back to the specs, total compute performance looks basically the same as the 4070 Ti Super: 44 TFLOPS FP32 and 352–353 TFLOPS FP16 on the tensor cores. Clock speeds are also lower, on paper, with the 5070 Ti. As usual, however, Nvidia's stated boost clocks are quite conservative estimates, and most of the games we tested ran at much higher speeds.

Memory speed and bandwidth are both quite a bit higher with the 5070 Ti. It's 33% faster than the 4070 Ti Super in speed and bandwidth and 78% higher in bandwidth compared to the 4070 Ti, thanks to having a 33% wider interface. And, of course, it also offers 33% more VRAM capacity than the 4070 Ti.

We tossed a few earlier GPUs from the 30- and 20-series into the table for reference. If you're still using something like a 3070 Ti or 2070 Super, the 5070 Ti should offer a pretty massive improvement in performance. Is it enough to get people to upgrade? For some, we're sure the answer will be yes, but if you're already sporting something like an RTX 4070 Ti Super or even an RTX 4070, this will likely be an easy generation to sit out.

Basically, you get improved AI features, including MFG support. You also get upgraded display outputs (DisplayPort 2.1 UHBR20) and a PCIe 5.0 x16 interface, though neither of those is likely to matter much for most consumer workloads. Power requirements are a bit higher than the 4070 Ti Super, with a TGP of 300W compared to 285W on the prior generation. (TGP stands for "Total Graphics Power," and it's what Nvidia calls the power of the entire graphics card, including all components. AMD uses the term TBP for Total Board Power, which ends up being mostly the same thing with a few minor nuances.)

You also get the much-disparaged 16-pin power connector, with a 3x 8-pin adapter, should you need it. This is the newer 12V-2x6 standard, but even that hasn't proven 100% reliable with the 50-series cards. The 5070 Ti shouldn't have any meltdown issues, considering it's only rated for 300W. That's better than the 5090, which seems to be following at least partially in the footsteps of the 4090. Pulling 600W through such a design just seems ill-advised, and the lack of more advanced monitoring functions hasn't helped matters.

Nvidia takes a staggered approach to GPU launches every generation, typically starting at the top and working its way down. The 5070 Ti is the third card from the Blackwell series, using the same GB203 GPU as the 5080 but with only 83% of the GPU cores enabled. Building redundancy and the ability to down-bin chips into a GPU design is common practice, so while the 5080 requires a basically perfect chip that's fully enabled, anything that can't meet the criteria can potentially be used in a 5070 Ti — and in the future, we'll undoubtedly see some 5070 or even 5060 Ti variants that use "leftover" GB203 chips.

Performance on paper may be up to 17% slower than the 5080, give or take, for 25% less money. And for games and tasks that are more memory dependent, the gap may only be 5~10 percent. At lower resolutions and settings, everything starts to become CPU and platform-limited, but we don't think most people buying a $750 to $1,000 GPU will want to run such low settings.

Let's check out the Asus RTX 5070 Ti Prime in more detail before hitting the benchmarks.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Asus GeForce RTX 5070 Ti Prime

For this review, Nvidia provided us with an RTX 5070 Ti Prime from Asus. It sports reference clocks and also has a base $749 MSRP. However, we don't know if this is "special launch pricing" like Asus did with certain 5090 and 5080 models, so prices might go up after the initial batch — and probably will, but we'll discuss that more in the conclusion. Still, this isn't a factory overclocked model, and we expect most base MSRP 5070 Ti cards will offer similar performance.

There are no Founders Edition RTX 5070 Ti cards, just as Nvidia made no 4070 Ti or 4070 Ti Super cards. This will be a GPU served solely by the AIB partners, which is probably a good thing for those businesses. The margins may be lower than on the 5080 and 5090, but we anticipate more people will be buying the 5070 Ti than either of the higher-tier parts.

As a base model RTX 5070 Ti, Asus doesn't provide much in the way of visible extras with the Prime. You get a triple-fan card, a 3x 8-pin to 16-pin adapter, velcro strips, and a quick-start manual. There's no support stand or similar hardware, which could still be beneficial on a card that weighs 1180g (2.60 lbs).

The 5070 Ti Prime is a 2.5-slot design measuring 304x126x50 mm. The fans, meanwhile, are custom 88mm designs with an integrated rim to improve airflow and static pressure. The center fan also spins in the opposite direction, which Asus says reduces turbulence and noise levels. The back fan (furthest from the video ports) also features a flow-through design to improve cooling by allowing the air to pass straight through the radiator fins.

The default clock speed on the GPU is a 2452 MHz boost clock, but if you install Asus GPU Tweak III and enable its OC mode, you get a 30 MHz boost. Which is basically something you could do — and a lot more — on your own without any difficulty. We haven't had time to fully test overclocking on the 5070 Ti, and we mostly don't worry too much about such things as modern GPUs tend to push pretty close to their reasonable limits, but if it's like the 5080, there should be a bit more headroom than usual.

Higher-spec Asus cards often include three DisplayPort connections and two HDMI connections, but the Prime sticks with the more traditional triple DP and single HDMI. There's also no lighting of any form on the card, RGB or otherwise. It's a no-frills design, intended to compete on price by eschewing any extra bells and whistles. That doesn't mean it's a completely base-level card, however.

Asus does have a VBIOS switch to toggle between quiet and performance modes. Even Nvidia still skips that on its reference cards. We tested in performance mode, but if you want a slightly quieter experience, graphics cards with a Q/P toggle often adjust the fan speed curves and GPU clocks slightly to reduce noise levels.

Asus also has some less visible extras under the hood. It has moved to phase-change thermal pads with all of its graphics cards now. These tend to hold up better over time compared to traditional TIMs (thermal interface materials) that can dry out. Asus also applies an adhesive, which it calls GPU Guard, to the corners of the GPU package. This can help to reduce warping and cracking over time, leading to better longevity.

Overall, the Asus 5070 Ti Prime follows the traditional approach for GPUs, with clean lines and nothing to draw attention to the card aesthetics. The shroud has a rounded edge on the top and a beveled edge on the bottom, and it's mostly black with a few silver accents. If you prefer a minimalist look to your graphics card, the Prime should be a good choice.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Nvidia RTX 5070 Ti Test Setup

We've revamped our test suite and our test PC, wiping the slate clean and requiring new benchmarks for every graphics card we want to have in our GPU benchmarks hierarchy. That will take time, and as with the 5090 and 5080 reviews, we don't have a ton of cards in this review. We've added the RX 7900 XT as well, but for now, we've got mostly Nvidia cards.

The Nvidia Blackwell RTX 50-series GPUs also bring some new technologies that require separate testing. Chief among these (for gamers) is the new DLSS 4 with Multi Frame Generation (MFG). That requires new benchmarking methods, and we want to spend some additional time with some DLSS 4-enabled games to get a better idea of how they look and feel. As with DLSS 3 frame generation, MFG isn't a magic bullet that makes everything faster and better. It adds latency, and the experience also depends on the GPU, game, settings, and monitor you're using.

MFG can potentially double the number of AI-generated frames (DLSS 4 can generate 1, 2, or 3 depending on the setting you select), smoothing out the appearance on your display. However, user input gets sampled only on the rendered frames, so MFG in 4X mode running at "160 FPS" would only be sampling the mouse and keyboard at 40 FPS. Our experience so far is that there's a minimum FPS that's needed for MFG to feel playable, and a higher level FPS where it starts to feel smooth and responsive.

Our GPU test PC has an AMD Ryzen 7 9800X3D processor, the fastest current CPU for gaming purposes. (We also tested the RTX 5090 on our old 13900K test bed, so check the prior review for those results. Suffice it to say that, at 4K, things are much closer than at 1080p.) We also have 32GB of DDR5-6000 memory from G.Skill with AMD EXPO timing enabled (CL30) on an ASRock X670E Taichi motherboard.

We're running Windows 11 24H2, with the latest drivers at the time of testing. We used AMD's 24.12.1 drivers, Nvidia's preview 571.86 drivers for the RTX 5090, 4090, and 4080 Super, slightly newer 572.02 drivers for the RTX 5080, and 572.43 preview drivers for the 5070 Ti, 4070 Ti, and 4070 Ti Super. (We also retested a few things on the 5090/5080 using these drivers, like 3DMark's DXR feature test that was previously running poorly.)

At some point in the coming days, once the new GPU launches have calmed down a bit, we'll see about retesting all of the new 50-series GPUs with up-to-date drivers. We anticipate optimizations and tuning will help the 50-series quite a bit in the next couple of months.

Our PC is hooked up to an MSI MPG 272URX QD-OLED display, which supports G-Sync and Adaptive-Sync, allowing us to properly experience the higher frame rates that GPUs like the 5080 and 5090 can deliver (especially with MFG). Most games won't get anywhere close to the 240Hz limit of the monitor at 4K when rendering at native resolution, which is where framegen and MFG can be useful.

The new GPU test suite consists of 22 games. We're still looking at some potential changes and additions, but this is where we're at for now. Six of the games in our standard test suite have RT support enabled. The remaining 16 games are run in pure rasterization mode. However, with cards like the RTX 5080 and 5090, we'll be looking at supplemental testing using some of the most demanding games that feature full RT / "path tracing" support. (That testing is still ongoing, but check page six.)

All 22 games are tested without any upscaling or frame generation as our baseline. Again, we plan to do additional investigations into things like DLSS 2/3/4 and framegen/MFG, but that's separate from the primary testing. There are noticeable differences between the image quality of DLSS, FSR, and XeSS, as well as differences in how much they can affect performance, which is why we're not using any of them for our baseline measurements.

All games are tested using 1080p 'medium' settings (the specifics vary by game and are noted in the chart headers), along with 1080p, 1440p, and 4K 'ultra' settings. This provides a good overview of performance in a variety of situations. Depending on the GPU, some of those settings don't make as much sense as others, but seeing how fast cards like the RTX 5080 run at 1080p can be enlightening. (As with our 5090 initial testing, 1080p performance seems to hit bottlenecks on the 5080 that are lower than what we see on the previous generation, likely due to driver immaturity for the new Blackwell architecture.)

Our OS has all the latest updates applied. We're also using Nvidia's PCAT v2 (Power Capture and Analysis Tool) hardware, which means we can grab real power use, GPU clocks, and more during our gaming benchmarks. We'll cover those results on page eight.

Finally, because GPUs aren't purely for gaming these days, we run some professional and AI application tests. We've previously tested Stable Diffusion, using various custom scripts, but to level the playing field and hopefully make things a bit more manageable (AI is a fast moving field!), we're turning to standardized benchmarks. We use Procyon and run the AI Vision test as well as the Stable Diffusion 1.5 and XL tests; MLPerf Client 0.5 preview for AI text generation; SPECworkstation 4.0 for Handbrake transcoding, AI inference, and professional applications; 3DMark DXR Feature Test to check raw hardware RT performance; and finally Blender Benchmark 4.3.0 for professional 3D rendering.

At present, Procyon does not work on the Blackwell GPUs due to an older and apparently outdated TensorRT framework.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Nvidia RTX 5070 Ti Rasterization Gaming Performance

We're dividing gaming performance into two categories: traditional rasterization games and ray-tracing games. We benchmark each game using four different test settings: 1080p medium, 1080p ultra, 1440p ultra, and 4K ultra.

For the RTX 5070 Ti, the 1440p ultra results might be the most interesting. They're useful in their own right, but they're also a stand-in for 4K with quality mode upscaling. 1440p upscaled to 4K will run a bit slower due to the overhead of DLSS, and you can also look at the balanced and performance modes if you want higher framerates, so the 1080p ultra results are also useful.

As a $750 card, the 5070 Ti can generally handle 1440p and 4K, particularly with a bit of upscaling help on the latter. We also have the overall performance geomean, the rasterization geomean, and the ray tracing geomean. Just to keep things easier to parse, we're going to put the charts in order from highest resolution to lowest on each group of charts.

We'll start with the rasterization suite of 16 games, as that's arguably still the most useful measurement of gaming performance. Plenty of games that have ray tracing support end up running so poorly that it's more of a feature checkbox than something useful.

We'll provide limited to no commentary on most of the individual game charts, as the numbers speak for themselves. The Geomean charts will be the main focus, as they provide the big picture of how the 5070 Ti compares to other GPUs.

There are two primary comparison points for the RTX 5070 Ti: the RTX 4070 Ti that launched in January 2023, and the RTX 4070 Ti Super that replaced the original card in January 2024. The expectation is that we'll see a bigger generational improvement over the vanilla 4070 Ti, while the 4070 Ti Super will be within striking distance of the 5070 Ti (unless you want to include MFG performance).

Rasterization performance tends to be higher so we can even look at the 4K results without reaching beyond the capabilities of the 5070 Ti. It averages just over 60 fps at 4K ultra across our test suite, beating the 4070 Ti by 25% overall. Drop to 1440p and the gap shrinks a bit to 18%, then 13% at 1080p ultra, and 11% at 1080p medium. That's not too surprising as even a $750 GPU can be CPU limited at 1080p.

The comparison to the 4070 Ti Super is decidedly less impressive. The 5070 Ti takes a 14% lead at 4K, where its additional bandwidth proves most beneficial. The lead decreases to 11% at 1440p, and 7–8 percent at 1080p. It's a small enough difference that, outside of running benchmarks like we're doing here, most people wouldn't be able to tell the two GPUs apart based on these results.

It's also worth noting that there are some games where the 4070 Ti Super comes out ahead of the 5070 Ti. Based on what we know of the architectures and specifications, this shouldn't happen unless the drivers aren't fully tuned for the new Blackwell GPUs. There are some results in our testing where the 4070 Ti Super comes out ahead of the 5070 Ti, and even a handful of cases where the 4070 Ti claims a victory. As we've previously noted with the RTX 5080 and 5090 cards, it seems Nvidia has more tuning worth to get done to properly leverage the new features in Blackwell chips.

One final point of comparison is AMD's RX 7900 XTX. Originally a $999 part, it dropped as low as $800 during sales before the supply dried up. AMD has done better in rasterization performance against Nvidia's GPUs, and that holds true here as well, though the margins aren't really meaningful in this case. The 7900 XTX beats the 5070 Ti by a few percent across all tested settings, with individual games going to one card or the other depending on the engine and other factors.

It's basically a tie, with AMD having a slight edge in performance on a card that's no longer worth buying since it now starts at around $1,350. So unless it comes back in stock at more reasonable prices, AMD's enthusiast RDNA 3 GPU ties Nvidia's new high-end card.

Below are the 16 rasterization game results, in alphabetical order, with short notes on the testing where something worth pointing out is present.

Assassin's Creed Mirage uses the Ubisoft Anvil engine and DirectX 12. It's also an AMD-promoted game, though these days, that doesn't necessarily mean it always runs better on AMD GPUs. It could be CPU optimizations for Ryzen, or more often, it just means a game has FSR2 or FSR3 support — FSR2 in this case. It also supports DLSS and XeSS upscaling.

Baldur's Gate 3 is our sole DirectX 11 holdout — it also supports Vulkan, but that performed worse on the GPUs we checked, so we opted to stick with DX11. Built on Larian Studios' Divinity Engine, it's a top-down perspective game, which is a nice change of pace from the many first-person games in our test suite.

Black Myth: Wukong is one of the newer games in our test suite. Built on Unreal Engine 5, which supports full ray tracing as a high-end option, we opted to test using pure rasterization mode. Full RT may look a bit nicer, but the performance hit is quite severe. (Check our linked article for our initial launch benchmarks if you want to see how it runs with full RT enabled. We've got supplemental testing coming as well.)

Dragon Age: The Veilguard uses the Frostbite engine and runs via the DX12 API. It's one of the newest games in my test suite, having launched this past Halloween. It's been received quite well, though, and in terms of visuals, I'd put it right up there with Unreal Engine 5 games — without some of the LOD pop-in that happens so frequently with UE5.

Final Fantasy XVI came out for the PS5 last year, but it only recently saw a Windows release. It's also either incredibly demanding or quite poorly optimized (or both), but it does tend to be very GPU limited. Our test sequence consists of running a set path around the town of Lost Wing.

We've been using Flight Simulator 2020 for several years, and there's a new release below. But it's so new that we also wanted to keep the original around a bit longer as a point of reference. We've switched to using the 'beta' (eternal beta) DX12 path for our testing now, as it's required for DLSS frame generation, even if it runs a bit slower on Nvidia GPUs.

Flight Simulator 2024 is the latest release of the storied franchise, and it's even more demanding than the above 2020 release — with some differences in what sort of hardware it seems to like best. Where the 2020 version really appreciated AMD's X3D processors, the 2024 release tends to be more forgiving to Intel CPUs, thanks to improved DirectX 12 code (DX11 is no longer supported).

God of War Ragnarök released for the PlayStation two years ago and only recently saw a Windows version. It's AMD promoted, but it also supports DLSS and XeSS alongside FSR3. We ran around the village of Svartalfheim, which is one of the most demanding areas in the game that we've encountered.

Hogwarts Legacy came out in early 2023, and it uses Unreal Engine 4. Like so many Unreal Engine games, it can look quite nice but also has some performance issues with certain settings. Ray tracing, in particular, can bloat memory use, tank framerates, and also causes hitching, so we've opted to test without ray tracing. (At maximum RT settings, the 9800X3D CPU ends up getting only around 60 FPS, even at 1080p with upscaling!)

Horizon Forbidden West is another two years old PlayStation port, using the Decima engine. The graphics are good, though I've heard at least a few people think it looks worse than its predecessor — excessive blurriness being a key complaint. But after using Horizon Zero Dawn for a few years, it felt like a good time to replace it.

The Last of Us, Part 1 is another PlayStation port, though it's been out on PC for about 20 months now. It's also an AMD-promoted game and really hits the VRAM hard at higher-quality settings. Cards with 16GB or more memory do just fine, in general.

A Plague Tale: Requiem uses the Zouna engine and runs on the DirectX 12 API. It's an Nvidia-promoted game that supports DLSS 3, but neither FSR nor XeSS. (It was one of the first DLSS 3-enabled games as well.) It has RT effects, but only for shadows, so it doesn't really improve the look of the game and tanks performance.

Stalker 2 is another Unreal Engine 5 game, but without any hardware ray tracing support — the Lumen engine also does "software RT" that's basically just fancy rasterization as far as the visuals are concerned, though it's still quite taxing. VRAM can also be a serious problem when trying to run the epic preset, with 8GB cards struggling at most resolutions. There's also quite a bit of microstuttering in Stalker 2, which framegen can help smooth out. (UE5 games tend to benefit more from framegen, in my experience, than some other games.)

Star Wars Outlaws uses the Snowdrop engine, and we wanted to include a mix of options. It also has a bunch of RT options, though our tests don't enable ray tracing. As with several other games, turning on maximum RT settings in Outlaws tends to result in a less than ideal gaming experience, with stuttering and hitching.

Starfield uses the Creation Engine 2, an updated engine from Bethesda, where the previous release powered the Fallout and Elder Scrolls games. It's another fairly demanding game, and we run around the city of Akila, one of the more taxing locations in the game.

Wrapping things up, Warhammer 40,000: Space Marine 2 is yet another AMD-promoted game. It runs on the Swarm engine and uses DirectX 12, without any support for ray tracing hardware. We use a sequence from the introduction, which is generally less demanding than the various missions you get to later in the game but has the advantage of being repeatable and not having enemies everywhere.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Nvidia RTX 5070 Ti Ray Tracing Gaming Performance

The RTX 5070 Ti should, if anything, do even better in games with ray tracing enabled. Nvidia was the driving force behind the creation of dedicated ray tracing hardware, which influences the DirectX Raytracing (DXR) and Vulkan Ray Tracing APIs. It all started with the Turing RTX 20-series GPUs, with each subsequent generation doubling the ray/triangle intersection calculation rates (per RT core).

Most RT games end up being better optimized for Nvidia GPUs, because Nvidia has been pushing the tech far more than AMD or Intel. We've selected six reasonably demanding RT games for our testing, and we'll add additional supplemental RT / full RT / DLSS 4 testing on the next page.

The RTX 5070 Ti mostly delivers the expected gains in ray tracing performance, with the exception of Minecraft where there's a driver bug of some form that severely impacts performance — particularly at 1080p and 1440p "ultra" settings. (It doesn't happen on our older 13900K testbed, so it seems to be related to AMD's 9800X3D and socket AM5 platform in some way.) Control also bumps into its 240 FPS limit at 1080p medium settings, which brings down the overall average.

Overall, the 5070 Ti offers up to 25% higher performance on average compared to the 4070 Ti, but only up to 11% higher performance than the 4070 Ti Super. It's also up to 30% faster than the 7900 XTX. And that's with the sometimes underperforming Minecraft results.

Toss the Minecraft data out and the 5070 Ti sees a more consistent 9–13 percent performance improvement over the 4070 Ti Super, 13–24 percent better performance than the 4070 Ti, and 18–22 percent higher framerates than the 7900 XTX. (Minecraft can help or hurt, depending on the resolution and settings, as it otherwise pushes a lot of RT effects and usually runs much better on Nvidia GPUs.)

Outside of Minecraft, these games only do modest amounts of RT. We'll be checking some full RT games on the next page (in the future) to show how the 5070 Ti Blackwell GPU fares against its progenitors in a heavier RT scenario.

Here are the combined rasterization plus ray tracing overall results charts. These charts use the geomean of all 22 games we tested, with RT accounting for slightly more than a quarter of the overall score.

Looking at the wide-angle view of GPU performance this way, the 5070 Ti does come out slightly ahead of AMD's 7900 XTX. It's only by 3–5 percent overall, though, so it's still basically a tie on performance. Except, the Nvidia card costs less, uses less power, and has some extra features that AMD's GPU lacks.

Meanwhile, the RTX 4070 Ti and 4070 Ti Super stories don't really change. The 5070 Ti delivers the largest gains at 4K ultra, where it's up to 25% faster than the 4070 Ti and 13% faster than the 4070 Ti Super.

The individual RT gaming charts follow, again with limited commentary on each.

Avatar: Frontiers of Pandora uses ray tracing, but it's not particularly forthcoming on when and where it's used. Reflections, in general, don't appear to use RT, which is one of the most noticeable upgrades RT can provide. Instead, it's used for shadows and possibly global illumination and some other effects. What I can say for sure is that nothing in the menus (other than "BVH Quality") directly mentions ray tracing, and the performance hit doesn't seem to be as severe as in some games. Still, since there's RT of some form, this one gets lumped into our DXR suite.

If you want a game where ray tracing is both clearly visible and actually makes the game look better, without totally destroying performance, look no further than Control. It's now five years old, and we're using the Ultimate version, but it's still arguably the best example of using RT well. And probably a lot of that is because you're running around the Federal Bureau of Control, an office space of sorts that has good reasons to have plenty of glass windows that reflect the scenery.

Note that Nvidia's RTX 50-series GPUs have some rendering errors in Control right now, and there's a hard 240 FPS cap that impacts the 1080p results. (This game is on my chopping block if I decide I want to trim down the number of tests I'm running.)

Possibly the most hyped-up use of RT in a game, Cyberpunk 2077 launched with more RT effects than other games of its era, and later, the 2.0 version added full path tracing and DLSS 3.5 ray reconstruction. Ray reconstruction ends up looking the best but only works on Nvidia GPUs, so, as with upscaling, it can be a case of trying to compare apples and oranges.

We're using medium settings with RT lighting at medium and RT reflections enabled, and then the step up uses the RT-Ultra preset. In all cases, any form of upscaling or frame generation gets turned off. However, we'll have more details on Cyberpunk 2077 with RT-Overdrive on the next page. It's also currently the only game with a public release that supports DLSS 4.

F1 24 enables several RT effects on the ultra preset but leaves them off on medium. But then 1080p medium runs at hundreds of frames per second, so we went ahead and turned all the RT effects on for our testing. We use the Great Britain track for testing.

Minecraft supports full path tracing, as well as DLSS 2 upscaling on RTX cards. We don't enable DLSS, and the game doesn't even allow it on the RTX 50-series GPUs right now. Apparently, it has some sort of hard-coded check for an RTX 20-, 30-, or 40-series GPU is our best guess. Or it's just a driver bug of some form.

The 50-series GPUs also underperform in Minecraft by quite a lot — especially at 1080p, less so at 1440p. Nvidia is aware of the problem and presumably working on a fix. But we've been saying that for over a month now.

Last on our list of RT-enabled games, Spider-Man: Miles Morales doesn't look as nice with RT turned on as the previous Spider-Man: Remastered. The reflections are less obvious, and perhaps performance is better as a result. But beyond the RT effects, maxing out the settings in Miles Morales definitely needs more than 8GB of VRAM, and even 12GB cards like the 4070 Ti can struggle at times.

One final ray tracing benchmark we have is the 3DMark DXR Feature Test, where we report the average FPS rather than the calculated score. This is similar to full RT in a game, only done via a standalone benchmark and perhaps in a more vendor-agnostic fashion. Nvidia has also fixed a bug here that was causing Blackwell 50-series GPUs to underperform. We've retested the 5090 and 5080 this round.

The results, as expected, have AMD's fastest GPU at the bottom of our test group. It was already slower than the 4070 Ti, so there was no reason to expect it to beat the 5070 Ti.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Nvidia RTX 5070 Ti Full RT and DLSS 4 Testing (Coming later...)

Testing the RTX 5070 Ti in games that support DLSS 4 requires some extra effort. We ran short on time, and we're already looking forward to the RTX 5070 and AMD RX 9070 XT and 9070 GPUs. But we're going to run some tests similar to our DLSS 4, MFG, and full RT 5080 and 5090 benchmarks as soon as we can. We'll discuss things in more detail once we have more numbers. Until then, here's the short recap of Multi Frame Generation. It's a key element in Nvidia's performance claims for the entire Blackwell family of GPUs, and the results — if taken purely at face value — can be highly misleading.

Above, you can see a video in which Bryan Catanzaro, Nvidia's VP of Applied Deep Learning Research, discusses the changes to DLSS 4. Some changes only affect the RTX 50-series GPUs while others apply to all RTX cards. DLSS upscaling is one that gets changed for everything with a new Transformers algorithm.

The short summary is that DLSS Transformers are more computationally intensive but yield better image quality results. We've played around with Cyberpunk 2077 and feel there's more to this than just marketing hype. Even with the Balanced and Performance modes, DLSS Transformers can still look very good at 4K. It could actually deliver on Nvidia's "better than native" claims for DLSS.

What about Multi Frame Generation is a different can of worms. As the name implies, it generates multiple frames rather than interpolating just one. It can run in three modes: 2X (generate a single frame), 3X (generate two frames), and 4X (generate three frames). In principle, it's a lot like the original framegen, just with more generated frames.

It takes two input images plus motion vectors and depth buffers and works to generate high-quality intermediate frames. In practice, there are some big changes compared to DLSS 3 on the RTX 40-series. MFG runs off the Blackwell tensor cores. We don't know if that means it can leverage the FP4 number format, but that would make sense. The OFA in the 40-series was a fixed function unit, and Nvidia apparently has reached the point where it can do better framegen via a software algorithm running a new AI model than by the old OFA. The new tensor-based framegen also runs faster than the OFA variant, even on an RTX 40-series GPU (or at least on the RTX 4090).

There are around 75 games and applications that are DLSS 4 enabled, according to Nvidia, either with native support or via Nvidia App overrides. (Nearly all of them are via the Nvidia App right now.) You can use the new DLSS Transformers model to get improved image quality, with a slight hit to performance relative to the previously existing DLSS CNN (Convolutional Neural Network) models but with higher image quality. The performance hit does vary by the generation of hardware, however, and we expect transformers will run quite a bit slower than the CNN model on the 30- and 20-series RTX cards.

MFG now has native support in several games, including the latest 2.2.1 Cyberpunk 2077 patch, Alan Wake 2, and Hogwarts Legacy. We still need to test the 5070 Ti (and previous generation cards it replaces), so we'll hold off on any charts or analysis right now. But what we know of MFG from the 5080 and 5090 suggests it's not going to be a universal win.

In our view, MFG basically needs at least 30 FPS for rendered performance, and 40+ FPS is better, with MFG either doubling, tripling, or quadrupling that. But it does necessarily scale perfectly, often getting more like 75% better for each additional generated frame. At the same time, delays are introduced, and depending on the game, user, and controller of choice, the final result can be better or worse than the non-framegen / non-MFG experience.

The resulting output on your monitor generally looks smoother, but it doesn't necessarily feel that much better. The 5080, for example, got 36 FPS with 4K RT-Overdrive and DLSS Quality Transformers. Turning on MFG bumped that to 66, 95, and 122 FPS, with input sampling happening at 33, 32, and 31 FPS, respectively. That means responsiveness got worse at each level while the frame smoothing got better.

What happens with a slower GPU like the 5070 Ti? The gains will be smaller and in some cases a game might feel like it's running worse than before, even though the reported FPS is higher. So, check back and we'll update this section with more details.

More to come....

Nvidia RTX 5070 Ti: Content Creation, Professional Apps, and AI

Modern GPUs like the RTX 5070 Ti aren't just about gaming. They're used for video encoding and professional applications, and increasingly, they're being used for AI. We've revamped our professional and AI test suite to give a more detailed look at the various GPUs. We'll start with the AI benchmarks, as those tend to be more important for a wider range of users.

Procyon has multiple AI tests, and we've run the AI Vision benchmark along with two different Stable Diffusion image generation tests. The tests have several variants available that are all determined to be roughly equivalent (in output) by UL: OpenVINO (Intel), TensorRT (Nvidia), and DirectML (for everything, but mostly AMD). There are also options for FP32, FP16, and INT8 data types, which can give different results. We tested the available options and used the best result for each GPU.

Unfortunately, Procyon needs to be updated to work on the RTX 50-series GPUs, so we'll skip commenting on these tests for now. The 5070 Ti, 5080, and 5090 all fail to run any of the TensorRT workloads we've tried. Nvidia is aware of the problem, but we're not sure how long it will take for a fix to become available.

ML Commons' MLPerf Client 0.5 test suite does AI text generation in response to a variety of inputs. There are four different tests, all using the LLaMa 2 7B model, and the benchmark measures the time to first token (how fast a response starts appearing) and the tokens per second after the first token. These are combined using a geometric mean for the overall scores, which we report here.

While AMD, Intel, and Nvidia are all ML Commons partners and were involved with creating and validating the benchmark, it doesn't seem to be quite as vendor-agnostic as we would like. AMD and Nvidia GPUs only have a DirectML execution path, while Intel has both DirectML and OpenVINO as options. Intel's Arc GPUs score quite a bit higher with OpenVINO than with DirectML.

The RTX 5070 Ti delivers a good result here, beating the previous generation RTX 4080 Super as well as the 4070 Ti Super and 4070 Ti. It's 17% faster than the 4070 Ti Super, and 45% faster than the 4070 Ti. MLPerf doesn't use FP4 for these tests, so the additional memory bandwidth from the GDDR7 likely helps performance quite a bit.

AMD's 7900 XTX takes longer on the time to first token, but after that delay it does reasonably well; the 5070 Ti ends up being 15% faster, however.

We'll have some additional SPECworkstation 4.0 results below, but there's an AI inference test composed of ResNet50 and SuperResolution workloads that runs on GPUs (and potentially NPUs, though we haven't tested that). We calculate the geometric mean of the four results given in inferences per second, which isn't an official SPEC score but it's more useful for our purposes.

The RTX 5070 Ti again edges past the RTX 4080 Super in this test, and it leads the 4070 Ti Super by 18% and the 4070 Ti by 30%. That's on the higher end of the performance leads we saw in our gaming benchmarks. It's also 30% faster than the 7900 XTX.

For our professional application tests, we'll start with Blender Benchmark 4.3.0, which has support for Nvidia Optix, Intel OneAPI, and AMD HIP libraries. Those aren't necessarily equivalent in terms of the level of optimizations, but each represents the fastest way to run Blender on a particular GPU at present.

The RTX 5070 Ti only offers modest improvements over the previous generation. It's 8% faster than the 4070 Ti Super and 23% faster than the vanilla 4070 Ti. AMD's 7900 XTX falls quite a bit behind here, with the 5070 Ti delivering 84% higher performance in Blender.

SPECworkstation 4.0 has two other test suites that are of interest in terms of GPU performance. The first is the video transcoding test using HandBrake, a measure of the video engines on the different GPUs and something that can be useful for content creation work. We use the average of the 4K to 4K and 4K to 1080p scores. Note that this only evaluates speed of encoding, not image fidelity.

The RTX 5070 Ti has the same video encoding block as the RTX 5090 and RTX 5080, so performance should be the same. And it is. Minor differences in clock speeds as well as run to run variations can account for the ~1% difference between the Blackwell GPUs. There's a minor improvement in performance compared to the 40-series cards as well, while AMD's 7900 XTX delivers the highest speed. However, our previous GPU encoding testing also showed that AMD had lower image fidelity.

Our final professional app tests consist of SPECworkstation 4.0's viewport graphics suite. This is basically the same tests as SPECviewperf 2020, only updated to the latest versions. (Also, Siemen's NX isn't part of the suite.) There are seven individual application tests, and we've combined the scores from each into an unofficial overall score using a geometric mean.

The RTX 5070 Ti falls behind everything except the RTX 4070 Ti models this time. Curiously, the 4070 Ti even manages to beat the 4070 Ti Super, which shouldn't normally occur. AMD's drivers for its consumer cards tend to be more friendly toward these professional applications, which is why the 7900 XTX claims top honors.

These AI and professional tests are ultimately just one aspect of GPU performance, and if you only care about gaming they shouldn't exert much influence on your choice of GPU. That's especially true of the professional tests. AI could become something useful even for gaming, maybe, but higher Blender performance will only matter if you're actually using Blender for 3D modeling.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Nvidia RTX 5070 Ti: Power, Clocks, Temps, and Noise

All our gaming tests are conducted using an Nvidia PCAT v2 device, which allows us to capture total graphics card power, GPU clocks, GPU temperatures, and some other data as we run each gaming benchmark. We have separate 1080p, 1440p, and 4K results for each area.

The Asus RTX 5070 Ti Prime comes in well below its 300W TGP rating (Total Graphics Power), though it's only 12W shy of that mark at 4K ultra. It uses about 15 to 30 watts more power than the prior generation 4070 Ti, and pretty much the same power use as the 4070 Ti Super. Notice also how much more power AMD's 7900 XTX uses, even at lower resolutions. It basically goes full tilt, even in situations where it should be CPU limited. Check the table at the bottom of this page for the individual game power results.

Clock speeds among the different GPUs and architectures aren't super important, but it's interesting to see where things land. The official boost clock on the RTX 5070 Ti is supposed to be 2452 MHz, but the Asus Prime often breaks 2.8 GHz during our tests. On average, it had the highest clocks of any of the tested GPUs at 1080p, and the second highest clocks at 1440p and 4K — with the Gigabyte 4070 Ti taking the top spot.

Like the clock speeds, comparing GPU temperatures without considering other aspects of the cards doesn't make much sense. One card might run its fans at higher RPMs, generating more noise while being "cooler." So these graphs should be used alongside the noise and performance results.

Typically, most recent graphics cards try to stay at or below about 70C. The Asus RTX 5070 Ti Prime averaged 66C at 4K, and that drops to 62C at 1080p medium. What's really going on is that the fans spin more slowly (or perhaps not at all in some cases) when the GPU isn't being taxed.

We check noise levels using an SPL (sound pressure level) meter placed 10cm from the card, with the mic aimed right at the center of one fan: the center fan if there are three fans, or the right fan for two fans. This helps minimize the impact of other noise sources, like the fans on the CPU cooler. The new noise floor of our test environment and equipment is around 34 dB(A), due to the noise from the CPU cooling pump.

The Asus 5070 Ti Prime ends up being reasonably quiet, compared to these other high-end cards, but it's not able to match the 4070 Ti Super — which is also an Asus card, but it's from the higher tier TUF Gaming line. The TUF Gaming has larger fans, which means it doesn't have to spin them as fast to move the same amount of air as the Prime card.

Here's the full table of testing results, with FPS/$ calculated using the various launch MSRPs for the cards. That's because street pricing can fluctuate, and the prior generation RTX 40-series and RX 7000-series are basically no longer readily available — just as the new 50-series GPUs are out of stock. Latency results are included for some of the games as well, and you can see the game-by-game power figures.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Nvidia RTX 5070 Ti: A solid high-end graphics card

Nvidia's RTX 5070 Ti deserves plenty of accolades. It delivers solid high-end performance, taking over from where the 4070 Ti Super left off. It's not revolutionary, but at least it's (generally) faster and cuts the price by $50 — in theory. There's still work to be done by Nvidia on the drivers, however, as there's really no good reason why the 4070 Ti Super and even the slower 4070 Ti should, at times, beat the new 5070 Ti.

While the more expensive RTX 5080 felt disappointing for only offering minor performance improvements over the existing 4080 / 4080 Super, and for sticking with 16GB of VRAM, the 5070 Ti can get away with 16GB by virtue of costing $749 — again, in theory. It's only about 10–15 percent faster than its immediate predecessor, but it's also 20–30 percent faster than its direct namesake. And it has some extra stuff that the prior generation lacks.

Part of the difficulty with Nvidia's latest GPUs is that the names have shifted upward. The xx70-class GPUs at one point cost around $300–$400. Then they became $599 and even $799 parts. Now the 5070 Ti walks that back slightly with a $749 base MSRP. In a sense, it's actually carrying on from the $699 RTX 3080 and the $649 GTX 1080 Ti. Sure, the number has changed, but Nvidia has been trying to stretch the range of GPUs to much higher price segments and has changed the nomenclature as it sees fit.

Nvidia RTX 5070 Ti pricing: Will it ever be $749?

Let's talk about that pricing a bit more. Officially, the RTX 5070 Ti launches with a base MSRP of $749. That seems pretty decent, not that we wouldn't all want cheaper, faster, and better GPUs. But Nvidia has several problems it faces right now, some mostly of its own making, others from external factors. Let's start with the former.

There's a large backlog of people wanting to upgrade their graphics card. I remain firmly convinced that the original plan for Blackwell and the RTX 50-series GPUs was for the cards to begin launching last fall. It's not just about the "every two years, basically" release cadence that Nvidia has been using for about a decade. It's also about the dwindling supply of the RTX 40-series GPUs.

Every time a company releases new products, it has to clear out inventory of the existing product lines. This makes sense as presumably your new stuff is better than the old stuff — otherwise why make it? The difficulty is getting the timing as well as supply transition just right. Stop making the old products too late and you end up with a glut of parts, graphics cards in this case, that no one wants to buy. Conversely, if you halt manufacturing too soon, you end up with a void in the supply chain.

Looking at the past few months, it's pretty clear what happened this round. Blackwell was supposed to come out last year, all the wafer starts for Ada Lovelace RTX 40-series GPUs were phased out to clear the way ... and then Nvidia ran into a packaging problem on its data center Blackwell B200 parts. Most if not all of the original shipment of GB200 data center GPUs ends up in the scrap heap and needs to be restarted, and TSMC only has so much capacity to go around — and Nvidia is already using a lot of that capacity.

So, Nvidia has to push back the consumer Blackwell GPUs, which use the same process node and manufacturing lines as the data center GB200. And it can't order more Ada wafers because those also use the same wafer lines. Which means we end up with most of the previous generation RTX 40-series GPUs being sold out, prices go up, and everyone is champing at the bit waiting for the RTX 50-series to arrive.

That supply and demand imbalance still persists right now, which means the RTX 5070 Ti will almost certainly sell out fast, just like the 5080 and 5090. But while the 5090 offers a clearly new level of performance and features like 32GB of VRAM, the 5070 Ti is basically a cheaper, slightly slower replacement for the 4080 Super with some new features like MFG and FP4. It should sell well, but anyone that wanted a fast Nvidia card and was willing to spend $1,000 or more should have already purchased an RTX 4080 or 4080 Super in the past two years.

The other problem with pricing is that the MSRP seems to have been established with little consideration to new and upcoming tariffs. The new U.S. administration imposed a 10% tariff on imports from China in January, and another 10% tariff just came into effect. There are also 25% tariffs on imports from Mexico and Canada. In short, many or perhaps even all graphics cards will now have higher base costs.

That will have a cascading effect on retail prices, and certainly in the near term it doesn't look like there's any way to avoid that. 20% higher tariffs could directly equate to 20% higher prices, which means the "real" MSRP for the RTX 5070 Ti — as well as the 5080, 5090, and upcoming 5070, 9070 XT, and 9070 — should probably be 20% higher. If you don't like the $749.99 MSRP, $899.99 will feel even worse.

What we're getting at is that, short-term, RTX 5070 Ti prices are likely to shoot up from the base $749 MSRP. That's unfortunate but not really something we can stop, other than refusing to pay more and waiting. Long-term, however, the performance isn't there to support extreme prices, and the pent-up demand should eventually get cleared out with prices returning to normal. It might take several months, maybe even a year or more, but perhaps at some point in the future the RTX 5070 Ti will actually see widespread availability at $749. Or perhaps tariffs will keep all prices significantly higher than in the past.

We'll also need to see how AMD's upcoming RX 9070 XT stacks up, because frankly I think AMD might have expected a bigger bump in performance from the 5070 Ti. If it targeted "4070 Ti Super plus 30%" with its top RDNA4 part, it can now choose to either undercut Nvidia and gain market share, or increase its own price. Again, we need to wait and see.

RTX 5070 Ti Multi Frame Generation and final thoughts

On paper, the RTX 5070 Ti strikes a good balance between performance, features, and value. It's still an expensive high-end card, but it's certainly a better value than the RTX 5090 and RTX 5080. It's also not faster (most of the time) than the previous generation RTX 4080, at least not unless you want to factor in MFG — and perhaps you should.

Frame generation tends to be a polarizing topic, with Nvidia acting like it's the same as normally rendered frames. At the other extreme are the "never framegen" people who act like it has completely ruined every game that uses the technique. The reality falls somewhere in between.

MFG is not a bad option to have, is how we view it. On the right games, with the right settings, it can make them look and feel better. Sometimes it breaks, and you need to tweak some other settings to get the desired result, but again, It's not bad to have options.

MFG is one more tool in Nvidia's bag of tricks, and it can be helpful in the right situations. It's just not universally better in all situations. It also tends to work and feel better when the baseline performance is sufficiently high. If the final performance is only 100 FPS, meaning a 25 FPS input sampling rate with MFG 4X, that will often feel worse than a native 40 FPS to most people. Which is precisely the problem with MFG. If it's "2.5X faster" than 40 FPS but actually doesn't feel better, you know there's a disconnect.

So, who is Nvidia targeting with the RTX 5070 Ti? People with an RTX 3070 to 3080 (or lower) GPU who want to upgrade will find plenty to like. It will be about 50% faster in raw performance, and the new features can make it feel like more of a step up than that. At least there are no glaring flaws with the product other than concerns with availability and the possibility of scalpers spoiling the party. But if you already have an RTX 40-series GPU, you should give this generation a pass until something truly compelling comes out.

We also need to see what actual pricing and availability look like. At $749, the RTX 5070 Ti represents a reasonable high-end graphics card worth purchasing. If the price climbs to $899 or more, however, it becomes far less compelling. We’ve heard there will be more 5070 Ti cards at launch than all the 5090 and 5080 cards that have been sold so far, combined. But there are no concrete numbers, and Nvidia has a tradition of selling out on just about every new GPU generation. [Update: Yes, it sold out almost immediately. Additional supplies remain a question mark.]

The 5070 Ti will likely keep that trend going for at least the first month or two of its existence. Beyond that, we wait and see. Our overall view of the 5070 Ti comes primarily from its performance and suggested pricing, with the understanding that it will take time for supply and prices to stabilize. It's a good card at $749, and the AIBs are going to try to milk it for all its worth. If the price ends up being $899 and up for a while? Don't buy it, simple as that.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

If you’re looking for value in a high-performance gaming monitor, check out Asus’ ROG Strix line. These monitors typically offer near or the same performance as the more expensive Swift screens and can save you a few bucks. You still get the same rugged build quality and check-this-out styling, with slick LED lighting and glowing ROG logos.

The newest example I've encountered is the XG27ACDNG. This mouthful of letters describes a 27-inch Quantum Dot OLED running at QHD 2560x1440 resolution, with a 360 Hz refresh rate, Adaptive-Sync, HDR400, and a wide color gamut that fully earns the moniker “reference level.” Let’s take a look at what is one of the best gaming monitors currently available.

Asus ROG Strix XG27ACDNG Specs

If you want maximum color in any display, Quantum Dots should be on your checklist. This technology increases the coverage of OLED panels to around 110% of DCI-P3. The XG27ACDNG is the first example I’ve measured at over 111%. That color is also incredibly accurate. Asus calibrates each one before shipping and includes a data sheet in the box. My review sample measured so well that I could make no improvement. It nailed every number for gamma, color temp, and gamut. And it delivered this goodness in both SDR and HDR modes.

The XG27ACDNG supports HDR10 and has a VESA DisplayHDR 400 TrueBlack certification. With the Uniform Brightness feature turned off, I saw around 260 nits for a full white field pattern, 427 for a 25% window, and in HDR mode, a 3% window just cracked 1,000 nits. Coupled with the perfect black levels that OLEDs are known for, this means superlative contrast and an image depth unmatched by any other technology.

If you’ve read my past OLED reviews, you know that they are the best choice for high-performance gaming. The XG27ACDNG rises above the 240 Hz rank and file with a 360 Hz refresh rate and perfect video processing. Moving objects are just as crystal clear as static ones, and input lag is perceptually non-existent. Additional options include ELMB up to 180 Hz (instead of Adaptive-Sync) and OLED Anti-Flicker, which addresses the occasional stuttering one sees when using a variable refresh rate. It will also accept signals up to 3840x2160 pixels at 120 Hz.

The XG27ACDNG delivers all the visual appeal of an ROG display with futuristic styling and solid build quality. The back features LED lighting, and the front features a glowing logo to guide your hand to the OSD joystick and control keys. Asus also offers a Display Widget app that puts the OSD on the Windows desktop. You also get a rugged stand with a threaded mount at the top for mics and webcams. USB ports and a KVM function support peripherals.

The XG27ACDNG has everything you need to dominate your game world with a current street price of around $700. It doesn’t cut corners in any area and once you see the test results, you’ll want one.

Assembly and Accessories

I’ve seen a few Asus monitors packed in recyclable pulp, but the XG27ACDNG came in traditional crumbly foam. The panel snaps onto a two-part stand that assembles with a captive bolt. The accessory bundle comes in a nice, zippered bag and you get a slick VESA mount adapter for use with aftermarket arms. Cables include IEC power, two USBs, DisplayPort and HDMI. You also get a cleaning cloth and a sheet of ROG stickers to adorn your gaming PC.

Product 360

The XG27ACDNG puts most of its styling cues around the back. Up front, it’s all business. There’s a thin flush bezel and the bottom trim is quite narrow as well. A glowing ROG logo adorns a small protrusion in the center, which has an OSD joystick and two control keys behind it. The screen has a glossy outer layer, which isn’t too bad at rejecting reflection, but avoiding sunny windows is always a good idea with any OLED.

You can see a large ROG logo in the back made of dots and backlit. The LEDs can be controlled from the OSD and light up in any color with a selection of breathing, flashing or static effects. Smaller ROG graphics are on the beefy stand which has a small cable hole at the bottom. There’s 5/20 degrees tilt and 45-degree swivel along with 120mm of height adjustment and a 90-degree portrait mode. Movements are smooth and firm, with no play or wobble. At the top is a quarter-inch threaded socket for things like mics and webcams. The XG27ACDNG is a premium build in every way.

Other interesting bits include a phone slot in the base. It’s large enough to accommodate all but the biggest phones and hold them solidly. Just under the input panel are molded-in graphics to indicate different functions along with the mantra “For Those Who Dare.” You should dare to buy an XG27ACDNG because it’s awesome. Video inputs include two HDMI 2.1 and one DisplayPort 1.4. The USB-C port mimics DP functions. There are also USB-A and B ports, one upstream and three down, which are supported by a KVM feature in the OSD. A 3.5mm jack provides headphone audio. There are no internal speakers.

OSD Features

The XG27ACDNG has a packed OSD with every conceivable image and gaming option available. Pressing the joystick brings it up, and it’s divided into nine sections.

The fun starts with Gaming, and here you’ll find toggles for Variable Refresh Rate and ELMB. They can’t be used together and ELMB is only available at 180 Hz or below. It doesn’t include variable pulse width and it reduces brightness by around half. But it’s unnecessary because at 360 Hz with AS turned on, motion resolution is perfect. You would only consider using ELMB below 150fps.

This menu also has GamePlus which includes frame counters, aiming crosshairs with a dynamic option, sniper mode, timers and a stopwatch. The dynamic crosshair changes color on the fly to maintain contrast with the background. The picture modes are under the GameVisual heading and number nine in total. Racing is the default and best choice. It’s spot on in every metric and needs no adjustment. If you have trouble seeing in your game’s dark places, there are three shadow boost levels plus a dynamic option.

The Uniform Brightness option lives in the Image menu, with brightness and contrast and the four HDR modes. If it’s turned on, the peak white level is fixed at around 260 nits. Off, it jumps to 427 (25% window). You’ll have to change brightness to compensate, but I noticed that the XG27ACDNG retained separate values for each mode so once you’ve dialed it in, you can toggle Uniform Brightness off and on at will.

In the Color menu, you can choose between three color spaces. Wide Gamut is fine for everything but if you want sRGB, it’s here too. It is a better option than the sRGB Cal picture mode because it retains all adjustments. You also get fixed color temps with a user option and a selection of gamma presets.

OLED Care options are extensive and include compensation for logos, taskbars and graphical boundaries. You can turn on a pixel shift and run cleaning routines. It’s good to see this level of attention to detail but in my experience, there is no need for concern. I’ve been using an OLED monitor for almost a year now and it shows no signs of burn-in. It runs a cleaning routine every few days.

The XG27ACDNG has PIP and PBP options with the ability to view two video sources at once. For the LED lights in the back, Aura RGB includes five different effects and options for all colors. Aura Sync matches the light show to what’s happening on the screen.

Three of the joystick directions are programmable for quick access to often-used functions. You can also save your settings to two memories for later recall. Finally, you’ll find the KVM options in the System Setup menu where you can bind USB ports to video inputs for control of multiple systems with a single set of input devices.

Asus ROG Strix XG27ACDNG Calibration Settings

The XG27ACDNG does not require calibration in its Racing mode and I could not make any improvement when I tried. It’s that precise. The only choices one might make are brightness and color space. To set brightness, turn Uniform Brightness on, find your preferred light level, then repeat the process with UB turned off. The XG27ACDNG retains the brightness value for both scenarios, so you don’t have to readjust when switching back and forth. This is extremely handy. I’ve included commonly used output levels for both situations below.

When HDR signals are detected, the XG27ACDNG unlocks four new modes, all of which are reasonably accurate. TrueBlack is the best choice according to my measurements. If you want to tweak brightness and contrast, there’s a toggle that lets you do that.

Gaming and Hands-on

Though there isn’t much to setting video processing options on any OLED, it’s worth checking out the provided features, and the XG27ACDNG has a few. There’s no overdrive, but there is ELMB. It won’t work with Adaptive-Sync turned on and it maxes at 180 Hz. It has no artifacts, but brightness is reduced by half. There is no adjustment, it’s either on or off. If you have enough processing power to keep frame rates above 200fps, it’s better to stick with Adaptive-Sync. The XG27ACDNG proved compatible with both FreeSync and G-Sync in my tests.

A feature unique to Asus is OLED Flicker. Or more precisely, OLED Flicker compensation. This option has medium and high settings and helps in low frame rate situations when you might see a little stutter. If you stay above 200fps, you won’t need it. I noted that switching between HDR modes or toggling on OLED Flicker switched the XG27ACDNG out of HDR mode necessitating a trip to the game’s menu to turn it back on. If you can run over 200fps consistently, stick with Adaptive-Sync and OLED Flicker off. My GeForce RTX 4090 equipped PC managed numbers consistently above 320fps with detail maxed in Doom Eternal’s Horde Mode.

Gameplay was a premium experience with perfect resolution in moving objects and panned backgrounds. There was never any smearing or breakup. Though I’ve played on a few monitors with lower input lag, the XG27ACDNG isn’t far behind the very best and to my perception, it is as quick as anything else I’ve tested. If you’re coming from a slower monitor, you’ll need to adjust your perception a bit as it is very precise. Movements start and stop exactly where you intend with no delay. It’s hard to imagine a better display for any type of game, from casual to frenetic.

Since it can accept Ultra HD signals up to 120 Hz, I tried this option in Windows and while gaming. I do not recommend using 3840x2160 for everyday tasks. Since it is not a multiple of the native 2560x1440 pixels, clarity is significantly reduced. Small text is very hard to read. Gaming at this resolution drops the frame rate to 120fps, introducing visible blur. I returned to QHD signals for the remainder of the review.

The XG27ACDNG is a superb everyday monitor for anything you can do with a PC. The image is stunning in SDR and HDR mode making it a great tool for graphics work. Photoshop images are richly detailed with perfect color. sRGB mode is just a couple of clicks away when you need it. Watching video is always awesome on an OLED. Movement is smooth with gorgeous color and contrast.

Some users might be concerned about resolution, and indeed, there are 4K OLEDs available at higher cost. But consider this: contrast is so good that the picture looks sharp even at 109ppi pixel density. You’ll have to practically press your nose on the screen to see any dot structure. High contrast always delivers higher sharpness. And QHD means faster frame rates in your favorite games, and lower cost of entry.

Takeaway: The XG27ACDNG is pretty much flawless for gaming, video and productivity. 27-inch flat panels are an ideal size for footprint, pixel density and price. QHD resolution delivers high frame rates and the super quick response OLEDs are known for. I can’t imagine anyone regretting this purchase.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

I have some very speedy OLEDs in my database and the quickest ones are here today. At 480 Hz is Asus’ flagship PG27AQDP. 360 Hz is covered by Philips’ 27M2N8500, MSI’s MPG271QRX and Alienware’s AW2725DF. At 240 Hz is Aorus’ FO27Q2. Is that unfair? You’ll soon see that it isn’t one bit.

Pixel Response and Input Lag

Click here to read up on our pixel response and input lag testing procedures.

Smooth motion is a given with any OLED running at 240 Hz or faster. Like the XG27ACDNG, they will all render moving objects with the same level of detail and clarity as stationary ones. 360 or 480 Hz isn’t better in that regard.

However, input lag is another story, and I’ve found that it doesn’t always go hand in hand with refresh rate. The XG27ACDNG is very quick with 15ms of total lag. But the Philips 27M2N8500 I reviewed recently still hold the all-time record at 10ms. And check out the Aorus. With just 240 Hz, it rocks the chart at 13ms. The PG27AQDP sits in the middle at 15ms. Don’t get me wrong, these are all very fast monitors with performance that most people can’t differentiate. But if you decide between the two Asus screens, the more expensive one won’t be faster even though it has a higher refresh rate.

Test Takeaway: Any OLED running at 240 Hz or more will be super smooth and provide instant response. Differences are easier to measure than to perceive. The XG27ACDNG delivers superb performance with a smooth feel and no perceptible control lag. When it comes to subjective evaluation, it doesn’t get better than this, or any fast OLED. It’s simply the best technology for gaming monitors.

Viewing Angles

Quantum Dot OLED monitors usually show a slight red tint in their 45-degree side photo. That is a byproduct of the extra layer which creates subtle polarization. The XG27ACDNG is like other QD OLEDs in this regard. This is not an issue because there is no brightness reduction nor is there a change in gamma. The image looks the same from all horizontal angles when showing full color content. The top view is solid too with just a small reduction in gamma and brightness.

Screen Uniformity

To learn how we measure screen uniformity, click here.

I have yet to encounter an OLED that doesn’t have visually perfect screen uniformity. As a premium technology, this isn’t surprising. The XG27ACDNG is just under 5%, which is a very low score in my test. This is excellent performance.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

To read about our monitor tests in-depth, please check out Display Testing Explained: How We Test PC Monitors. We cover brightness and contrast testing on page two.

Uncalibrated – Maximum Backlight Level

The XG27ACDNG‘s peak brightness depends on a couple of things. If you turn on Uniform Brightness, the peak level stays the same regardless of screen content and tops out around 260 nits. If you turn CB off, the peak reaches 427 nits when measuring a 25% window pattern. Smaller windows are even brighter, up to 1,000 nits from a 3% pattern. You’ll need different brightness settings to equalize normal content but the XG27ACDNG retains those values which is convenient. The black levels can’t be measured so contrast is theoretically infinite.

After Calibration to 200 nits

On page one, you’ll find the brightness values for 200 nits with Uniform Brightness on or off. Adjusting the light level does not affect contrast; it’s still unmeasurable. The ANSI checkerboard used to measure intra-image contrast is also not affected. The black squares can’t be measured because the pixels are turned off. This is typical OLED performance, which is always excellent.

Test Takeaway: The XG27ACDNG delivers brightness and contrast typical of the 27-inch Quantum Dot OLED monitors I’ve tested. Uniform Brightness can be used to tailor the monitor to different environments and uses. Gaming benefits from it being off so you get punchier highlights. But productivity is less tiring when it’s turned on. This is a feature shared by most OLEDs but the XG27ACDNG stands out in its ability to retain different brightness values when switching back and forth.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

The XG27ACDNG includes a factory calibration, which can be used in the Racing picture mode. If you want to tweak, there are color temps and gamma presets. You can also change the gamut volume between sRGB, DCI-P3, and Native, which is a little more colorful than P3.

Grayscale and Gamma Tracking

Our grayscale and gamma tests use Calman calibration software from Portrait Displays. We describe our grayscale and gamma tests in detail here.

The XG27ACDNG’s default grayscale chart is as close to perfect as I’ve ever seen. RGB tracking and gamma are so close to the reference line that you can barely see the traces. It truly doesn’t get better than this. However, no improvement was found when I attempted a calibration, so there are no additional results to show.

The sRGB chart is nearly flawless except for a bump in gamma at the 90% step which means it’s too dark at that level. It manifests as slightly muted highlights, but the effect is subtle. If you need sRGB accuracy, the XG27ACDNG provides it.

Comparisons

I charted the XG27ACDNG’s 0.50dE value against the other screens in both the pre- and post-calibration comparisons and it doesn’t get much better. Only the Philips managed to eke out first place after adjustment. Visually, there is no difference between any of the monitors.

The XG27ACDNG’s gamma is spot on as well. The range of values is just 0.06 and it rides 0.45% off the 2.2 reference with an actual value of 2.21. These errors are wholly invisible in actual content.

Color Gamut Accuracy

Our color gamut and volume testing use Portrait Displays’ Calman software. For details on our color gamut testing and volume calculations, click here.

The XG27ACDNG is a bit over-saturated like all Quantum Dot monitors. This is not a bad thing. With no visible hue errors and linear saturation tracking, there is nothing to complain about. Again, calibration cannot improve upon the result.

The sRGB chart is reference-level as well. I achieved this by changing the color space option in the Color menu rather than employing the sRGB Cal picture mode. Doing it that way means you can still adjust all image parameters like gamma and color temp. sRGB Cal limits changes to Brightness only.

Comparisons

The XG27ACDNG posts an impressively low color error of 1.68dE, but it should be noted that it is not the only screen that achieves this without calibration. I also left the Aorus and MSI screens at their defaults because calibration didn’t improve them either. It’s safe to say that this category is extremely color-accurate and often does not need adjustment for great color.

The XG27ACDNG’s Quantum Dot tech makes it the first OLED I’ve measured over 111% for color volume when referenced to DCI-P3. It also covers an ideal 101.7% of sRGB. Accuracy is high enough for critical apps in the sRGB and DCI-P3 realms. And it’s currently the most colorful OLED you can buy.

Test Takeaway: There’s not much to say about the XG27ACDNG’s color other than that it is supremely accurate and supremely saturated. It’s a standout monitor that will satisfy anyone with its stunning and colorful image.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

Our HDR benchmarking uses Portrait Displays’ Calman software. To learn about our HDR testing, see our breakdown of how we test PC monitors.

For the best HDR, it doesn’t get better than a QD-OLED like the XG27ACDNG. It doesn’t have the brightness of a high-end Mini LED, but that tech can’t deliver the deep black levels required for truly impactful imagery. The XG27ACDNG switches automatically and offers four HDR-specific picture modes with adjustment capability.

HDR Brightness and Contrast

The XG27ACDNG defaults to Gaming HDR mode, but I found better luminance tracking using the TrueBlack option. This doesn’t affect the peak output, which is just over 434 nits for a 25% window. Asus claims 1,000 nits for small highlights, and I have no reason to doubt this. If you want higher output, the PG27AQDP stands above the others. But for practical use, there is barely a difference between any of the six monitors.

Grayscale, EOTF and Color

The XG27ACDNG’s HDR grayscale tracking is visually perfect with no errors greater than 1dE which is impressive. The EOTF tracks closest to reference in the TrueBlack mode where it runs just a bit dark until meeting the line at 40%. Some shadow detail might be a little hard to see but this is a minor issue.

In the color test, the XG27ACDNG makes full use of its large gamut to create vivid and saturated HDR. The measured points are past their targets but in a linear progression so there is no loss of detail. Textures are crisp and clean in all areas of the image. And there are no visible hue errors.

The BT.2020 chart is impressive and shows red coverage of around 90%, green at almost 80% and blue hitting 95%. This is excellent performance.

Test Takeaway: The XG27ACDNG’s HDR performance is on par with what I saw in the SDR tests. Color is very saturated and spot-on accurate. Luminance tracking is close to reference so textures and detail is as clear as it possibly can be. This is an awesome monitor for HDR content and whether you watch video or play games, it will look amazing.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

With new speedy OLED gaming monitors appearing on almost a weekly basis, you’d think prices would drop more quickly. Sadly, this hasn’t happened but that doesn’t mean the category lacks value. The image quality and gaming performance offered is simply next level. The only thing other technologies can boast is higher brightness, but nothing can beat OLED’s black levels and contrast.

LCD can’t match OLED’s video processing smoothness and low input lag either. All the examples I’ve tested were super smooth and had no perceptible delays in control response. And you don’t need to run at dizzying frame rates to eliminate motion blur. 360 Hz is more than enough. The Asus ROG Strix XG27ACDNG stands above the 240 Hz rank and file with its 360 Hz and perfect Adaptive-Sync. And it offers the options of ELMB strobing and OLED Flicker compensation.

To this wonderful goodness, it adds some of the best color I’ve ever seen. Not only does it cover 111% of DCI-P3, making it the most colorful OLED I’ve yet tested, it’s unfailingly accurate as well, as in equal to or better than any professional screen. And it nailed those numbers right out of the box, no calibration needed. It’s so good that I couldn’t make even the smallest improvement.

So, you’ve got a stunning image, unparalleled gaming performance and perfect color. The Asus ROG Strix XG27ACDNG has no flaws of consequence. Does that make it expensive at $700? Compared to the broad scope of 27-inch gaming displays, yes. But you’re getting one of the best gaming monitors available and there are certainly more expensive examples that don’t perform as well. If you want the best in a 27-inch flat OLED, this one should be on your shortlist.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

Benchmarking software developer PassMark publishes the average results of all Windows PC tests across the globe every two weeks in a line graph. In line with what many enthusiasts might expect, the PassMark graph has always shown a consistent increase in processor performance year-on-year. However, for the first time since the company started keeping track in 2004, the average CPU mark score for desktop and laptop processors has dropped, with laptops dropping 3.4% year-over-year.

We see the biggest drop in laptop CPU performance results. PassMark recorded an average result of 14,632 across 101,316 samples last year. But, in 2025, the average score sat at an average of 14,130 points between 25,541 samples, decreasing the average score by 3.4%.

The average desktop PC result in 2024 netted 26,436 points for 186,053 samples. But for 2025, the average score currently sits at 26,311 points for over 47,810 samples — a 0.5% drop from last year. While that drop is small, we should only see a continued progression of faster performance.

We’ve also seen results for the top-performing CPUs, and it seems that we’ve basically reached a performance plateau, will little to no uplift in PassMark scores for the past three years for desktop chips and laptop CPUs. This happened after we received a massive uplift of 58.6% in the Top Desktop CPU benchmark scores in 2023, with the introduction of the AMD Ryzen Threadripper Pro 7995WX. The arrival of the AMD Ryzen 9 7945HX3D laptop CPU in the same year also delivered a 69.9% jump in average performance for laptops. However, a new desktop chip that will dethrone the Threadripper Pro 7995WX is yet to arrive, while Intel’s new Core Ultra 275HX has provided a measly 6.8% increase in performance points on mobile.

What’s quite baffling, though, is that AMD, Intel, and even Qualcomm have just released new desktop and laptop CPUs that should have increased performance levels, according to their marketing. While it’s true that the AMD Ryzen 9000 series (except for the new X3D chips) and Intel Arrow Lake Core Ultra 200S processors were a disappointment for many enthusiasts, these should have at least delivered a modest increase in performance, particularly in productivity applications.

Passmark itself mused on X (formerly Twitter) that it could be that people are switching to more affordable machines that deliver lower power and performance. Or maybe Windows 11 is depressing performance scores versus Windows 10, especially as people transition to it with the upcoming demise of the latter. We've certainly seen plenty of examples of reduced performance in gaming with some of the newer versions of Windows 11, particularly as Intel and AMD struggled to upstream needed updates into the OS.

At the moment, we don’t know the cause of this drop, but it could also be that we’re just in the first quarter of 2025. As more people get newer gear and run tests for the remainder of the year, this number could climb and reflect the performance of newer chips that arrived recently, or will arrive in the coming months.

PassMark also muses that bloatware could contribute to the sudden decline in performance, but that seems like a longshot. In the end, the decline could simply be due to an odd interaction between the benchmark itself and the latest versions of Windows 11.

The Intel Arc B570 picks up where the Arc B580 left off, namely with an even lower price point of $219. On paper, that's 12% cheaper, but it also comes with a 10% reduction in core counts, 3.5% lower clocks, and most importantly a 17% reduction in VRAM capacity and bandwidth. If you're trying to save money it might be worth considering, but the value proposition isn't as strong as the B580. It will compete with the best graphics cards of the rising generation, but only by virtue of very likely being the least expensive new GPU that we'll see in the coming year.

We've covered the Intel Battlemage architecture and the Arc B-series GPUs already. Now it's time to see how the B570 stacks up in real-world testing. Shaving $30 off the price while also cutting the memory by 2GB may not be the best solution for gaming or AI usage going forward. But let's start with the specifications.

As noted already, the Arc B5780 takes the same BMG-G21 core and trims a few functional units, with a lower core clock as well. The resulting card ends up with 13% less theoretical compute and 17% less memory bandwidth — and also 17% less memory capacity.

We haven't seen very many 10GB graphics cards over the year. There was the original RTX 3080, and more recently AMD's RX 6700 (non-XT), and that's basically it. We know there are quite a few modern games that can exceed 8GB of VRAM use, so the B570 may have a bit more wiggle room. However, lossless memory compression techniques in GPUs can also have an impact, so raw capacity isn't the final word.

Considering the Arc B580 ended up outperforming the prior generation Arc A770 by around 20%, we expect the new B570 to be slightly faster than the A770 as well. Except, higher resolutions where the extra VRAM capacity and bandwidth come into play will likely still favor the older GPU. Not that you'd really want to pick up an A770 16GB, considering they're now selling at $300 or more again.

Intel gives the Arc B570 a Graphics Clock of 2500 MHz, with a maximum boost clock of 2750 MHz. We noticed with the B580 that all the cards, including factory overclocked models, seemed to keep the maximum boost clock, and that appears to be the case with the B570 as well. Without manual overclocking, you'll get 2750 MHz peak performance, and in practice nearly every game and application we tested hit that clock speed.

Given the $219 MSRP, the Arc B570 will primarily face off against existing GPUs from AMD and Nvidia. Nvidia hasn't made a sub-$249 graphics card since the RTX 3050 8GB debuted in early 2022, three years ago. It now goes for $200, while supplies remain. There's also a more recent RTX 3050 6GB card as well, which we haven't tested, that sells for $170.

But we never particularly cared for the RTX 3050 cards. They were too slow for ray tracing to be a selling point, and in rasterization performance AMD's RX 6600 was clearly superior — it even competes with the higher spec RTX 3060 12GB. Our RX 6600 vs RTX 3050 GPU faceoff ended up being a clear victory for the RX 6600, mostly based on its superior performance. Despite being over three years old, the RX 6600 remains readily available, with prices starting at $190.

Time constraints (with holiday breaks and CES travels) meant we couldn't test every GPU we'd like to include for this review. We'll have the same cards as the Arc B580, plus a couple of additions (RTX 3060 and RX 6600). But before we get to the benchmarks, let's take a closer look at the ASRock Arc B570 Challenger OC we received for review.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

ASRock Arc B570 Challenger OC

Intel isn't making any reference Arc B570 cards, leaving that to its partners. It sent us an ASRock Arc B570 Challenger OC for this launch review, and we have some other B570 cards incoming... but those will probably have to wait until after we've cleared the RTX 5090 and RTX 5080 launches. Indications so far are that the various partner cards will all perform similarly, so it really comes down to pricing and aesthetics.

The primary attraction for the Battlemage GPUs will be their value proposition. Arc B580 is a great option for $249. At the current higher demand and backorder prices of $350 or more, though, it's nowhere near as interesting. The same goes for the Arc B570. It really needs to sell for $219, not $249 or $269.

We assume the ASRock Challenger OC will be a base MSRP model, as that's usually how ASRock approaches GPUs — the Phantom Gaming, Steel Legend, and Taichi brands are reserved for higher spec models with more RGB lighting. The Challenger B570 does have a small RGB strip, but otherwise it's a relatively barebones design.

ASRock's Arc B570 Challenger OC measures 248x130x42 mm and weighs 721g. It's a reasonably compact design in terms of the length, but it's taller than many lower tier cards. While it's a 2-slot card, there's also a metal backplate that makes it slightly thicker than some graphics cards and could preclude the use of an adjacent slot above the primary x16 slot (if you have a motherboard with such a configuration).

ASRock equips the Challenger with two older style 95mm custom fans, without the integrated rims that have become common on higher quality GPUs. We didn't encounter any issues with keeping the card cool and reasonably quiet, but these feel more like leftover fans from several years ago rather than quality components. You get what you pay for, in other words.

The Challenger comes with the usual complement of three DisplayPort 2.1 outputs and a single HDMI 2.1 port. We assume it's similar to the B580 LE in that one of the DP2.1 outputs supports UHBR13.5 (54 Gbps) while the other two are UHBR10, but we don't have a DP2.1 monitor for testing. Instead, we're using a 4K 240 Hz display that leverages DSC (Display Stream Compression) and runs on the older DP1.4a standard.

The B570 has an official TBP (Total Board Power) rating of 150W, and ASRock provides a single 8-pin PCIe graphics power connector. That can deliver up to 150W on its own, with an additional 75W provided by the x16 slot, so there should be ample margins for any potential overclocking.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

For 2025 (as well as the Arc B580 review), we've upgraded our GPU test PC and modernized our gaming test suite. The new system has an AMD Ryzen 7 9800X3D processor, the fastest current CPU for gaming purposes. We did test the B580 on our old 13900K test bed as well, and most of the results were basically the same as the 9800X3D — meaning the GPU speed is the limiting factor in most games.

We're running Windows 11 24H2, with the latest drivers at the time of testing. We used AMD's 24.12.1 drivers, Nvidia's 566.36 drivers, and Intel's preview 6256 drivers for the B570. We also retested the B580 with the publicly available 6256 drivers, while the Arc A770 used the 6319 drivers.

Note that these changes mean all the results from our GPU benchmarks hierarchy, while still valid for when they were run, need to be refreshed. We'll be working on a revised GPU hierarchy in the coming weeks, but it will be a bit before that's fully ready — we want at least all the current generation cards to be included, and it's no secret that both Nvidia Blackwell RTX 50-series GPUs and AMD RX 9000-series RDNA 4 GPUs are incoming.

Our PC is hooked up to a Samsung Odyssey Neo G8 32, one of the best gaming monitors around, allowing us to potentially experience some of the higher frame rates that might be available on the fastest GPUs. Most games can't get anywhere close to the 240 Hz limit of the monitor, especially not with budget to midrange hardware like the Arc B570 and its direct competition.

The new (revised since B580) GPU test suite consists of 22 games. We dropped Call of Duty Black Ops 6 from the suite due to frequent changes and some other oddities, and we're still looking at other potential changes, but this is where we're at for now. We've also toned down on ray tracing tests, mostly because outside of a few select games, it often seems to kill performance for debatable image quality upgrades. So, while more of the games have RT support, it's only enabled in six of the games — and even then, the visual upgrades are only really noticeable in three of the games. The remaining 16 games are run in pure rasterization mode.

All 22 games were tested without any upscaling or frame generation. We'll see about doing additional XeSS testing in the future, but trying to compare DLSS 2/3 (and soon 4), FSR 2/3, and XeSS performance without accounting for differences in image quality strikes us as a bad baseline way of measuring performance. Plus, we'd rather the default in games be native rendering, leaving upscaling and framegen as true performance boosting options — so you can break 120 fps or 144 fps, rather than just trying to get to 60 fps.

All games are tested using 1080p 'medium' settings (the specifics vary by game and are noted in the chart headers), along with 1080p, 1440p, and 4K 'ultra' settings. Some may wonder about the reasoning behind the selected settings, so let's quickly elaborate.

What we want to show with graphics cards is how performance scales. We include 1080p medium as a baseline "everything released in the past few years ought to handle this" setting. Then moving to 1080p ultra provides enough of a gap to be interesting — sometimes it's still only 10% slower, but other games it might be half as fast as medium settings. If we tested 1080p high instead, that's potentially one less useful piece of information.

Going beyond 1080p ultra, we don't want to change both the resolution and the settings, as there's going to be a lot of overlap between 1440p medium and 1080p ultra as an example. So we just test 1440p and 4K ultra, at least where it makes sense. And keep in mind that today's ultra is tomorrow's high, the next day's medium, and next week's low — except it's more like a year or so between each level.

The end result is that our tests will show both how GPUs run at comparable settings, where some designs may have shortcomings (e.g. insufficient VRAM or bandwidth), and provide ways for people to extrapolate how things would run at other settings. While we don't test 1440p or 4K at medium settings, if you check the 1080p medium to ultra scaling on a slower GPU from the same vendor, that should also apply (roughly) to a higher tier GPU at higher resolutions.

As we're in the process of retesting everything on our new PC and test suite, we're toning down the number of comparison points. The most direct competition for the B570 is a bit hard to pin down. The RTX 4060 and RX 7600 XT clearly cost more, but we'll include them as the "step up" options alongside the B580. The RX 7600 also costs a bit more, but it's fairly close. Then we drop to the RX 6600 and RTX 3050 as sub-$200 options that cost less than the B570. Besides those cards, we also have Nvidia's RTX 3060 12GB (which isn't so readily available these days), and we also tested the Arc A770 and A750.

The primary competition for the B570 ends up being older GPUs that are on their way out, and we don't really expect any new AMD or Nvidia GPUs to target the sub-$250 market. Maybe we'll be wrong, but we suspect the eventual RTX 5060 will likely cost $299 or more, and RX 9060 will likewise probably cost $299 or more. That at least gives Intel a clear win as the least expensive new graphics card. If you want to get an idea of where other GPUs might land, check out our full GPU benchmarks hierarchy — and then use the percentage increase in the hierarchy and apply that to the test data from this review.

Our test PCs are now running Windows 11 24H2, with all the updates applied. We're also using Nvidia's PCAT v2 (Power Capture and Analysis Tool) hardware, which means we can grab real power use, GPU clocks, and more during our gaming benchmarks. We'll cover those results on the page with power use.

Finally, because GPUs aren't purely for gaming these days, we've run some professional and AI application tests. We've previously tested Stable Diffusion, using various custom scripts, but to level the playing field we're turning to standardized benchmarks. We use Procyon, and run the AI Vision test as well as the Stable Diffusion 1.5 and XL tests; MLPerf Client 0.5 preview for AI text generation; SPECworkstation 4.0 for Handbrake transcoding, AI inference, and professional applications; 3DMark DXR Feature Test to check raw hardware RT performance; and finally Blender Benchmark 4.3.0 for professional 3D rendering.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

We're breaking down gaming performance into two categories: Traditional rasterization games, and ray tracing games. Each game has four test settings, though for the B570 we're largely going to ignore the 4K ultra results. We also have the overall performance geomean, the rasterization geomean, and the ray tracing geomean.

We'll start with the rasterization suite of 16 games, because that's arguably still the most useful measurement of gaming performance, particularly for a budget GPU like the Arc B570. Each game has four charts, ordered by how we would rate their importance. For the B570, the order will be 1080p ultra, 1080p medium, 1440p ultra, and (just for laughs) 4K ultra.

Our overall rasterization results pretty much inform our opinion of the Arc B570. It achieves performance parity with the old RTX 3060 12GB, with worse 1% low results (i.e. drivers most likely). It's a bit faster than the 3060 at 1080p medium, basically tied at 1440p ultra, and 4K ultra performance struggles at best.

Give the price, looking more closely at 1080p gaming makes sense. AMD's cheaper RX 6600 wins in a few games, but the B570 easily outclasses it overall. It's 29% faster than the 6600 at 1080p ultra, and 14% faster at 1080p medium. And if you care, it's also 45% faster at 1440p ultra and 34% faster at 4K ultra. Is that worth the additional $30 in price? We think so.

Against the RX 7600, performance is a bit of a wash. Arc B570 leads by 10% at 1080p ultra, but falls 10% behind at 1080p medium. It gets pyrrhic victories at 1440p and 4K as well, leading by 23–24 percent overall, if that matters to you. But in this case, Arc B570 costs $30 less (at the time of writing, assuming MSRP pricing on the B570), so that's another easy win.

But what about the Arc B580? It costs $30 more, in theory — prices are jacked up right now. That's 14% more money, for 18% more performance at 1080p ultra, 13% higher performance at 1080p medium, and it's 22% faster at 1440p ultra. 4K ultra really wants more than the B570's 10GB of VRAM, and so the B580 ends up being 51% faster there.

Below are the individual rasterization results, in alphabetical order with limited commentary.

Assassin's Creed Mirage uses the Ubisoft Anvil engine and DirectX 12. It's an AMD-promoted game as well, though these days that doesn't necessarily mean it always runs better on AMD GPUs. It could be CPU optimizations for Ryzen, or more often it just means a game has FSR2 or FSR3 support — FSR2 in this case. It also supports DLSS and XeSS upscaling.

The B570 falls well behind the RX 7600 here, at least at 1080p. It does win at 1440p, however, and still manages a very playable 68 FPS in that case. Upscaling could make 4K run fine as well.

Baldur's Gate 3 is our sole DirectX 11 holdout — it also supports Vulkan, but that performed worse on the GPUs we checked, so we opted to stick with DX11. Built on Larian Studios' Divinity Engine, it's a top-down perspective game, which is a nice change of pace from the many first person games in our test suite.

The Arc B570 (and B580) really underperform here. We speculated that it was related to drivers, but that's a month ago and after four new drivers we still haven't a significant performance improvement. Even the RX 6600 generally beats the B570 here.

Black Myth: Wukong is one of the newer games in our test suite. Built on Unreal Engine 5, with support for full ray tracing as a high-end option, we opted to test using pure rasterization mode. Full RT may look a bit nicer, but the performance hit is quite severe. (Check our linked article for our initial launch benchmarks if you want to see how it runs with RT enabled.)

The Battlemage GPUs continue to underperform in Black Myth Wukong. We don't have a clear explanation as to what's going on, but it's one of a few games where Intel's GPU drivers still need work.

Dragon Age: The Veilguard uses the Frostbite engine and runs via the DX12 API. It's one of the newest games in my test suite, having launched this past Halloween. It's been received quite well, though, and in terms of visuals I'd put it right up there with Unreal Engine 5 games — without some of the LOD pop-in that happens so frequently with UE5.

The B570 basically ties the RTX 3060 at 1080p, with only the 7600 XT and B580 beating it at 1440p (at least among the tested GPUs — there are lots of faster cards not included in these charts).

Final Fantasy XVI came out for the PS5 last year, but it only recently saw a Windows release. It's also either incredibly demanding or quite poorly optimized, but it does tend to be very GPU limited. Our test sequence consists of running a path around the town of Lost Wing.

None of the tested GPUs do very well in Final Fantasy XVI. The 7600 XT takes the top spot at 1080p and 1440p, but the B570 is only somewhat playable at 1080p ultra.

We've been using Flight Simulator 2020 for several years, and there's a new release below. But it's so new that we also wanted to keep the original around a bit longer as a point of reference. We've switched to using the 'beta' (eternal beta) DX12 path for our testing now, as it's required for DLSS frame generation even if it runs a bit slower on Nvidia GPUs (not that we're using framegen).

The B570 comes in slightly behind the RX 7600 at 1080p and 1440p, but ahead of the RTX 4060. It's interesting that the Nvidia GPUs don't do too well here, possibly because of DX12.

Flight Simulator 2024 tends to struggle on 8GB cards at the ultra setting, and the B570 comes out ahead of the 4060 again at 1080p. AMD's RX 7600 and 6600 really struggle here, except at 1080p medium.

God of War Ragnarök released for the PlayStation two years ago and only recently saw a Windows version. It's AMD promoted, but it also supports DLSS and XeSS alongside FSR3. We ran around the village of Svartalfheim, which is one of the most demanding areas in the game that we've encountered.

The B570 underperforms in this game, falling behind the RTX 3060 at all four settings (not that 4K is really viable). It's not VRAM capacity either, as the 4060 does fine at 1440p and below.

Hogwarts Legacy came out in early 2023, and it uses Unreal Engine 4. Like so many Unreal Engine games, it can look quite nice but also has some performance issues with certain settings. Ray tracing in particular can bloat memory use and tank framerates, so we've opted to test without ray tracing.

The B570 lands between the RTX 4060 and RTX 3060, and this is one of the games where Nvidia's older GPU takes the lead over the newer model. Chalk that up to VRAM capacity.

Horizon Forbidden West is another two years old PlayStation port, using the Decima engine. The graphics are good, though I've heard at least a few people that think it looks worse than its predecessor — excessive blurriness being a key complaint. But after using Horizon Zero Dawn for a few years, it felt like a good time to replace it.

The B570 lands closer to the B580 here than in many of the other games, though performance tanks hard at 4K ultra. Minimum FPS isn't that great on the B570 either, though the latest drivers did improve the B580 performance consistency.

The Last of Us, Part 1 is another PlayStation port, though it's been out on PC for about 20 months now. It's also an AMD promoted game, and really hits the VRAM hard at higher quality settings. The B570 struggles here at ultra settings, falling to the bottom of the charts. At medium settings, it comes out just behind the RX 7600 and ahead of the RTX 3060. It's interesting that the RTX 4060 doesn't seem to be bothered much by only having 8GB of VRAM.

A Plague Tale: Requiem uses the Zouna engine and runs on the DirectX 12 API. It's an Nvidia promoted game that supports DLSS 3, but neither FSR or XeSS. (It was one of the first DLSS 3 enabled games as well.)

The B570 trades blows with the older A770 across our test settings. It also swaps spots with the 4060 a few times. AMD's 7600 and 7600 XT do pretty well here for an Nvidia promoted game.

Stalker 2 is another Unreal Engine 5 game, but without any hardware ray tracing support — the Lumen engine also does "software RT" that's basically just fancy rasterization as far as the visuals are concerned, though it's still quite taxing. VRAM can also be a serious problem when trying to run the epic preset, with 8GB cards struggling at most resolutions.

The B570 trails the RTX 3060 at 1080p, takes a slight lead at 1440p, and then performance collapses at 4K. It's basically only sufficient for native 1080p, though, and needs upscaling for 1440p to be viable.

Star Wars Outlaws uses the Snowdrop engine and we wanted to include a mix of options. Outlaws also happens to be one of the games where we've had the most difficulty on Arc GPUs, with continued rendering errors on the B580 and B570 even with the latest drivers. Crashes still occur with Battlemage as well.

Despite the stability issues, the B570 does beat the RX 7600 XT and below from AMD, and performance isn't too far off the RTX 4060 at 1440p and below. Only 1080p is really playable without upscaling.

Starfield uses the Creation Engine 2, an updated engine from Bethesda where the previous release powered the Fallout and Elder Scrolls games. It's another fairly demanding game, and we run around the city Akila, which is one of the more taxing locations in the game.

The B570 struggles here again, just like the B580. It's mostly playable at 1080p ultra, but 1% lows are in the teens, with lots of micro-stuttering.

Wrapping things up, Warhammer 40,000: Space Marine 2 is yet another AMD promoted game. It runs on the Swarm engine and uses DirectX 12, without any support for ray tracing hardware. We use a sequence from the introduction, which is generally less demanding than the various missions you get to later in the game but has the advantage of being repeatable and not having enemies everywhere.

The B570 beats the RTX 3060 but trails the 4060 and RX 7600. It's also faster than the RX 6600, though, so as a budget option it's not too bad.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

As a $219 graphics card, ray tracing isn't really a major considering on the Arc B570. Yes, it can run some RT games fine, and it's generally much better with RT than AMD's existing GPUs. We'll have to see how RDNA4 changes things when it arrives.

Most RT games end up being better optimized for Nvidia GPUs, because Nvidia has been pushing the tech far more than AMD or Intel. We selected six reasonably demanding RT games for our testing (and we may add Indiana Jones and the Great Circle at some point in the future).

The Arc B570 does pretty well overall in our geomean for ray tracing. AMD's GPUs fall to the bottom of the charts, though 4K RT proves to be too much for the B570. It's basically on par with the previous generation A770 16GB, despite having far less theoretical compute and a lot less VRAM.

Nvidia's RTX 4060 beats the B570 at 1080p, but the card are basically tied at 1440p. Intel's new budget card also leads the 3060 at 1440p and below. If you're looking for a sub-$250 card that can sort of handle RT, the new Arc B570 might suffice, but in general you'll probably want to just leave RT disabled for this level of GPU.

And now that we have all the rasterization and ray tracing results, we can also look at the big picture. These charts use the geomean of all 22 games that we've tested, with RT basically accounting for a quarter of the overall score.

We think it's fair to say that there are a lot of RT games where the tech doesn't really do much other than tank performance, but there are also a select few games that definitely benefit visually from RT. So, we have far more rasterization games in our suite but still include a handful of RT games to give a more balanced overall view of how the GPUs stack up.

The Arc B570 ends up placing just ahead of the RTX 3060, and the ray tracing games really punish AMD's 8GB GPUs so that they fall to the bottom of the charts at ultra settings. The B570 and A770 16GB end up being pretty closely matched overall. But let's look at the individual games again.

Avatar: Frontiers of Pandora uses ray tracing, but it's not particularly forthcoming on when and where it's used. Reflections in general don't appear to use RT, which is one of the most noticeable upgrades RT can provide. Instead, it's used for shadows and possibly global illumination and some other effects. What I can say for sure is that nothing in the menus (other than "BVH Quality") directly mentions ray tracing, and the performance hit doesn't seem to be as severe as in some games. Still, since there's supposed to be RT of some form, this one gets lumped into our DXR suite.

The B570 does decently overall, though it's a bit odd how close the A750 gets, even taking the lead at 1440p and 4K. Probable VRAM bandwidth is a factor. For 1080p, the B570 leads everything except the RTX 4060 and B580, and it trades blows with the A770.

If you want a game where ray tracing is both clearly visible and actually makes the game look better, without totally destroying performance, look no further than Control. It's now five years old, and we're using the Ultimate version, but it's still arguably the best example of using RT well. And probably a lot of that is because you're running around the Federal Bureau of Control, an office space of sorts that has good reasons to have plenty of glass windows that reflect the scenery.

AMD's GPUs have a lot of issues with Control, especially on the 8GB cards. Performance degrades over time, which didn't happen a couple of years back. It's almost like AMD stopped worrying about the game at some point.

The Arc B570 is able to match or exceed the performance of the RTX 4060 in this game, not too shabby for a $220 GPU.

Possibly the most hyped up use of RT in a game, Cyberpunk 2077 launched with more RT effects than other games of its era, and later the 2.0 version added full path tracing and DLSS 3.5 ray reconstruction. Ray reconstruction ends up looking the best but only works on Nvidia GPUs, so as with upscaling it can be a case of trying to compare apples and oranges.

We're using medium settings with RT lighting at medium and RT reflections enabled, and then the step up uses the RT-Ultra preset. In all cases, any form of upscaling or frame generation gets turned off.

AMD's GPUs struggle badly in Cyberpunk 2077, but the Arc B570 mostly manages a playable level of performance at 1080p. 1440p is out of reach without upscaling, though the game does have XeSS support if you want to give that a shot.

F1 24 enables several RT effects on the ultra preset but leaves them off on medium. But then 1080p medium runs at hundreds of frames per second, so we went ahead and turned all the RT effects on for our testing.

The B570 results are odd here. B580 does quite well in F1 24, but the B570 falls well off the pace. It could be the reduced VRAM, or perhaps newer drivers have caused a drop in performance. The B570 mostly trails the RTX 3060 in this game, at least in our testing.

Minecraft supports full path tracing, and that kills performance on the AMD GPUs. The Arc B570 does better, but it's only reasonable for 1080p. Performance also dropped quite a bit with the B570 at 1440p and 4K, falling below the A750.

Last on our list of RT-enabled games, Spider-Man: Miles Morales doesn't look as nice with RT turned on as the previous Spider-Man: Remastered. The reflections are less obvious, and perhaps performance is better as a result. But beyond the RT effects, maxed out settings in Miles Morales definitely needs more than 8GB of VRAM.

The B570 does well at 1080p both at medium and maximum settings. 1440p starts to show a lot of performance inconsistency, however, and 4K performance died hard. So stick with 1080p if you want maxed out settings on the new Arc GPU.

One final ray tracing benchmark we have is the 3DMark DXR Feature Test, where we report the average FPS rather than the calculated score. This is similar to full RT in a game, only done via a standalone benchmark and perhaps in a more vendor agnostic fashion.

Things here are... odd. We retested the B580, as the B570 with newer drivers initially delivered better performance. The B580 did improve and is now 15% faster than the B570, right in line with expectations. That's good to see, but then we still have the A750 matching the B580 with the A770 performing 10% faster than the B580. Does that mean the A770 still has more RT performance? It might, or it might just be something where further tuning of drivers will help.

The RT hardware in Battlemage is supposed to be up to twice as fast as what was in Alchemist, so 20 ray tracing units in the B580, running at higher clocks, should be able to surpass the performance of the 32 RTUs in the A770. That it doesn't do so in this test raises some interesting questions, and we don't have all the answers just yet.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

Modern GPUs aren't just about gaming. They're used for video encoding, professional applications, and increasingly they're being used for AI. We've revamped our professional and AI test suite to give a more detailed look at the various GPUs. We'll start with the AI benchmarks on the Arc B570, as those tend to be more important for a wider range of users.

Procyon has several different AI tests, and for now we've run the AI Vision benchmark along with two different Stable Diffusion image generation tests. The tests have several variants available that are all determined to be roughly equivalent by UL: OpenVINO (Intel), TensorRT (Nvidia), and DirectML (for everything, but mostly AMD). There are also options for FP32, FP16, and INT8 data types, which can give different results. We tested the available options and used the best result for each GPU.

Arc B570 performs well in the Stable Diffusion tests, with the only faster GPU in our budget to midrange collection for SDXL being the B580. The 4060 does pull ahead in the older and less taxing SD 1.5 test, but the value proposition looks strong on the B570 for these AI workloads. Note that the SDXL test failed to run on the RX 6600, after multiple tries — it needs more VRAM, apparently.

The AI Vision test uses older workloads like ResNet50 to test performance, and again the B570 takes second place behind the B580. How applicable these results are in the real world remains debatable, but for properly optimized software Battlemage looks good.

ML Commons' MLPerf Client 0.5 test suite does AI text generation in response to a variety of inputs. There are four different tests, all using the LLaMa 2 7B model, and the benchmark measures the time to first token (how fast a response starts appearing) and the tokens per second after the first token. These are combined using a geometric mean for the overall scores, which we report here.

While AMD, Intel, and Nvidia are all ML Commons partners and were involved with creating and validating the benchmark, it doesn't seem to be quite as vendor agnostic as we would like. AMD and Nvidia GPUs only currently have a DirectML execution path, while Intel has both DirectML and OpenVINO as options. We reported the OpenVINO numbers, which are quite a bit higher than the DirectML results.

The B570 takes the second spot again, after the B580. All of the Arc GPUs have a fast time to first token result, which again seems to stem from the OpenVINO path. For the average tokens per second, Battlemage takes first and second, but AMD's 7600 XT ranks third, ahead of the 4060. The older RX 6600 falls far behind, both in time to first token and average tokens per second.

We'll have some other SPECworkstation 4.0 results below, but there's an AI inference test composed of ResNet50 and SuperResolution workloads that runs on GPUs (and potentially NPUs, though we haven't tested that). We calculate the geometric mean of the four results given in inferences per second, which isn't an official SPEC score but it's more useful for our purposes.

Again, software optimizations will likely make or break performance here. The A770 comes out on top, followed by the 4060 and 3060. All three of the AMD GPUs fall in the middle, then the A750 trails the A770 by a larger than normal margin, and the Battlemage GPUs take the two bottom slots.

For our professional application tests, we'll start with Blender Benchmark 4.3.0, which has support for Nvidia Optix, Intel OneAPI, and AMD HIP libraries. Those aren't necessarily equivalent in terms of the level of optimizations, but each represents the fastest way to run Blender on a particular GPU at present.

The B570 and Battlemage results once more fall behind some GPUs we would normally expect them to beat, like the A750 and A770. Nvidia's 4060 and 3060 place first and second, while AMD's GPUs bring up the rear. As this test should stress the hardware RT units, it could indicate that the reduced number of RT units (relative to Alchemist) is to blame, or it could be a lack of tuning in the current Intel drivers.

SPECworkstation 4.0 has two other test suites that are of interest in terms of GPU performance. The first is the video transcoding test using HandBrake, a measure of the video engines on the different GPUs and something that can be useful for content creation work. Here we use the average of the 4K to 4K and 4K to 1080p scores.

Here the Arc B570 does very well, matching the B580 and easily surpassing the other GPUs. You can also see the different generations of video codec hardware at play, with the two 7600 cards basically tied while the 6600 is quite a bit slower. Battlemage appears to have some improvements that help relative to Alchemist in this particular workload as well.

Again, it's disappointing that Intel killed off the studio portion of its drivers that allowed users to easily capture video content. OBS hasn't worked for me on the Battlemage cards, perhaps drivers again, or possibly just because it's more complex and I didn't set it up correctly.

Our final professional app tests consist of SPECworkstation 4.0's viewport graphics suite. This is basically the same tests as SPECviewperf 2020, only updated to the latest versions. (Also, Siemen's NX isn't part of the suite now.) There are seven individual application tests, and we've combined the scores from each into an unofficial overall score using a geometric mean.

The Arc B570 lands at the bottom of the overall chart, and does somewhat poorly relative to the other GPUs in several of the individual tests. 3dsmax, Creo, and Maya have the B570 in last place, often quite a bit behind the B580. AMD meanwhile offers a strong point for its GPUs in this set of tests, so if you use any of these professional applications, check the individual results.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

All of our gaming tests are conducted using an Nvidia PCAT v2 device, which allows us to capture total graphics card power, GPU clocks, GPU temperatures, and some other data as we run each gaming benchmark. We have separate 1080p, 1440p, and 4K results for each area.

The Arc B570 has an official power rating of 150W. That's for the whole board, not just the GPU — and we've noticed that some software seems to only grab GPU power (RivaTuner Statistics Server as an example). But the PCAT provides real power use.

As we saw with the B580, the B570 comes in below its rated power level. We say average power use across our test suite of 130–140 watts, slightly more than the RX 6600 but less than the RX 7600 — and far below the previous generation A750 and A770. But Nvidia's RTX 4060 still ranks as the lower power use of the tested GPUs.

Clock speeds among the different GPUs and architectures aren't particularly important, but it's interesting to see where things land. Intel specifies a "Graphics Clock" of 2500 MHz for the B570, and that's supposed to be an average across a suite of games. But it's a conservative average if so, and in fact the ASRock B570 card basically ran at it's maximum 2750 MHz boost clock in nearly all of our tests.

Like the clock speeds, comparing GPU temperatures without considering other aspects of the cards doesn't make much sense. One card could run its fans at higher RPMs, generating more noise while being "cooler." So these graphs should be used alongside the noise and performance results.

The ASRock B570 card ends up in the top three for running cool and quiet. It's slightly warmer than the RTX 4060 and RX 6600, landing just above 60C under load.

We check noise levels using an SPL (sound pressure level) meter placed 10cm from the card, with the mic aimed right at the center of one fan: the center fan if there are three fans, or the right fan for two fans. This helps minimize the impact of other noise sources like the fans on the CPU cooler. The noise floor of our test environment and equipment is around 31–32 dB(A).

[Charts to come, sorry! Still testing...]

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

It's nice to see a new graphics card that has a theoretically budget price of $219. We'll have to wait and see how things develop over time, though, as the Arc B580 has been sold out and overpriced basically since it launched last month. Hopefully things calm down soon and prices drop to MSRP levels.

Could the same thing happen to the Arc B570? Yes, but it probably won't be as severe. The B580 ends up as the more desirable card, since performance scales more than the price. For most graphics cards, you'll typically pay 20% more money for 10% more performance, rather than 12% more money for 18% more performance.

We haven't talked a lot about drivers this time, but as with our B580 testing, there are definitely anomalies and issues that still need fixing. Some games underperform, other crash on a regular basis. This is the Intel "fine wine" argument where performance and compatibility improve over time. But most people would likely prefer having something work properly from the start.

Battlemage in many ways feels like a rehash of the Alchemist launch. It's faster than the outgoing AMD and Nvidia GPUs, but new models are coming soon. But we don't expect Nvidia's future RTX 5060 to cost less than $300, and even AMD's RX 9060 may start at $300. That leaves the sub-$250 and "closer to $200" market to the Arc B570 by default.

It still has to contend with older model cards, but the RX 6600 isn't able to keep pace. Even in rasterization games, the B570 ends up being almost 30% faster than the 6600, while costing about 16% more. So unless prices on the RX 6600 drop another $20 or more, the B570 wins that comparison.

It also wins against the RX 7600. It's only slightly faster in general, and tied at 1080p medium, but it costs $30 less. Basically, the Arc B570 ends up standing alone as a brand new chip that only costs $220. There's nothing else released in the past two years that sits in that price bracket.

If you're mostly concerned with price and value rather than performance and compatibility, the Arc B570 should be a reasonable option. But if you can afford the extra $30 — and find a B580 in stock for that price — its sibling GPU remains the better option.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content

CES marks the beginning of a new year in tech, a time to see new products that make us excited about what’s to come in the 12 months ahead. Every January, we make the trek to Las Vegas with one goal: find and report on the most interesting, innovative and downright promising computers, components and peripherals.

This year’s show is an embarrassment of high-performance riches, with new high-end graphics cards announced by Nvidia and its partners, along with a first-of-its-kind AI workstation. But even more appealing are the products that allow humans to overclock their own abilities and experiences. These include a monitor that helps you keep track of your schedule / PC performance, a multicolor 3D printer and a laptop with a screen that expands to help you get more work done.

These are the 16 best products of CES 2025.CES marks the beginning of a new year in tech, a time to see new products that make us excited about what’s to come in the 12 months ahead. Every January, we make the trek to Las Vegas with one goal: find and report on the most interesting, innovative and downright promising computers, components and peripherals.

This year’s show is an embarrassment of high-performance riches, with new high-end graphics cards announced by Nvidia and its partners, along with a first-of-its-kind AI workstation. But even more appealing are the products that allow humans to overclock their own abilities and experiences. These include a monitor that helps you keep track of your schedule / PC performance, a multicolor 3D printer and a laptop with a screen that expands to help you get more work done.

These are the 16 best products of CES 2025.

Best Graphics Card: Asus ROG Astral RTX 5090 32GB OC

Asus ROG Astral RTX 5090 32GB OC

 Nvidia pulled the wraps off its RTX 50-series Blackwell graphics cards, with the 5090 starting at $1,999, 5080 at $999, 5070 Ti at $749, and 5070 at $549. All of the new GPUs look to offer new features, higher performance, and improved image quality — with AI playing a key role in all those areas. But the card we're most excited to see is the RTX 5090, and the Asus ROG Astral represents the top of the product stack.

While Nvidia has a Founders Edition 5090 coming in a dual-slot, dual-fan package, we're concerned about how hot that might end up getting in typical use. It's the same size as the old RTX 3080 Ti Founders Edition, more or less, a card that ran loud and hot with only a 350W TGP. 575W for the 5090 will need a lot of cooling, and maybe Nvidia's design with liquid metal will surprise us, but the safer bet will be a chunky card like the Astral OC.

It comes with four fans, three on the front and one on the back. Coupled with a big triple-slot cooler and a factory overclock, it should deliver some of the best out of the box performance we'll see from the coming generation. And the RTX 5090 looks to be a beast, with 170 Streaming Multiprocessors, 21,760 CUDA cores, 32GB of GDDR7 memory, and 1.8 TB/s of bandwidth. Oh, and it has x16 PCIe 5.0 support.

For workloads that are GPU limited, we anticipate the RTX 5090 will deliver at least a 25–30 percent boost in performance relative to the prior generation RTX 4090. For AI tasks and games that leverage the new features of the Blackwell architecture, it could be up to twice as fast. We'll have a full review when the product launches later this month.

Read More: Nvidia announces RTX 5090 at $1,999

—Jarred Walton

Best CPU: AMD Ryzen 9 9950X3D

AMD Ryzen 9 9950X3D 

AMD’s 16-core 32-thread Ryzen 9 9950X3D takes the company’s latest leading-edge technology to a new performance tier in productivity applications while still delivering among the best gaming performance on the market. AMD’s game-boosting 3D V-Cache technology, which uses an L3 cache chiplet to boost gaming performance, has left the company alone at the top of the gaming charts, easily beating Intel’s competing processors.

The X3D chips are so popular that the current best CPU for gaming, the Ryzen 9 9800X3D, remains almost impossible to find at retail. We expect a fresh supply of Ryzen 9 9950X3D chips to help alleviate some of those concerns, especially because AMD says it lands within one percent of the 9800X3D’s gaming performance. Naturally, it will come at a higher price point because it has more cores, but AMD will find no shortage of enthusiasts searching for the best of the best.

AMD released the Ryzen 9 9800X3D late last year with a new spin on the company’s cache tech — AMD now places the cache chiplet under the die instead of on top, thus providing more thermal headroom so the processor can deliver higher performance than the prior-gen models. Paired with the Zen 5 microarchitecture and the TSMC 4nm node, the new design delivers outstanding gaming performance. 

However, the X3D chips still have limitations that prevent them from providing the same level of performance in heavily-threaded work as standard ‘non-X3D’ processors. That’s where the 9950X3D’s extra cores, paired with the addition of a second die that lacks a cache chiplet so it can operate at higher clock rates, step in. The 9950X3D won’t match similar 16-core AMD chips that don’t have the extra cache chiplet, but this design will provide enough performance in heavily-threaded applications for all but the most demanding users.

The 9950X3D chips will launch in the first quarter of this year, but AMD hasn’t announced pricing yet. Intel remains entirely incapable of matching the X3D chips in gaming in any price range, so AMD will continue to dominate the gaming market throughout 2025.

Read More: AMD launches Ryzen 9 9950X3D and 9900X3D, claims 20% faster gaming performance than Intel’s flagship Arrow Lake processors

—Paul Alcorn 

Best Gaming Monitor: HP Omen 32x

HP Omen 32x smart gaming monitor

Lots of companies are bumping up their refresh rates, contrast ratios, or other specs here at CES 2025, but HP has gone in a different direction to deliver a monitor that is truly innovative and — at least to me as someone who’s used TVs as monitors for more than a decade — something deeply appealing. The 31.5-inch, 144 Hz Omen 32x is HP’s first monitor to include Google TV for all your streaming needs.

But as much as a gaming monitor that does double-duty as a compact streaming TV is appealing, you could always just buy a small TV and use it as a monitor or vice versa. What makes the Omen 32x stand out is that it’s designed to do both from the ground up, and HP has incorporated some extremely slick picture-by-picture setups where you can, say, easily stream a YouTube tutorial from your phone alongside a game, or catch up on your favorite TV show while you grind away in an MMORPG.

You could do some of that with a TV by fumbling clumsily with a remote, but HP’s software makes it look simple to jump between inputs and screen layouts, in a way that appealed to my multitasking / procrastinating brain like nothing else here at CES 2025. I wish the company had opted for an OLED screen rather than IPS. And I’d love to see a larger size (say, 42 or even 48 inches) to make the device as appealing as a TV as it is as a monitor, but there’s always next year.  

Read More: HP’s new Omen gaming monitors include 280 Hz IPS Black and Google TV options

 —Matt Safford

Best Gaming PC: Alienware Area-51 desktop

Alienware Area-51 desktop   

Not everyone has the know-how to build or upgrade a PC, which is why prebuilts are popular. And in terms of popularity, Alienware’s rigs have long held a spot near the top among those who would rather buy a PC than build one. But once you’ve invested in a high-end gaming rig like the new Alienware Area-51 desktop, chances are you’ll want to make it last as long as possible.

And it’s the novice-friendly, longevity-extending design that we like most about Dell’s latest gaming desktop flagship — or should that be mothership? Sure, it supports all the latest and greatest hardware, like Nvidia’s RTX 5090, PCIe 5.0 SSDs, and up to a 420mm radiator in its top chamber. But the company has moved to an ATX motherboard and power supply, making future upgrades much easier than with previous Alienware rigs. And the company has printed QR codes around key component areas like the PSU, GPU, CPU, and RAM, so buyers can easily bring up tutorials and upgrade advice videos from their phones if they have a hardware issue or want to upgrade.

Sure, you could still save money by building your own rig if you know what you’re doing and have the time to put together a PC. But as one of the most popular prebuilt desktop sellers, it’s nice to see Alienware doing something to keep gamers from just tossing their whole system and buying something new when they aren’t getting the frame rates that they want or a component fails outside of its warranty.   

Read More: New Alienware Area-51 laptops have a Gorilla Glass window, color-shifting paint job

 —Matt Safford

Best Laptop: Lenovo ThinkBook Plus Gen 6

Lenovo ThinkBook Plus Gen 6 Rollable laptop

You can never have too much screen real estate, particularly vertical screen real estate, which helps with productivity and multitasking. However, when you need to carry a laptop around, anything larger than a 14 or 15-inch display can get quite bulky in your bag. Enter Lenovo’s ThinkBook Plus Gen 6, the first laptop with a screen that expands with the touch of a button.

Using rollable display technology, the ThinkBook Plus Gen 6 grows from 14 to 16.7 inches and increases its screen resolution from 2000 x 1600 to 2000 x 2350. You can trigger this expansion either by pressing a button in the upper right corner or the keyboard or by using hand gestures. Then, when it’s time to pack up, just shrink the screen again before closing the lid.

To help you take advantage of this added pixels, Lenovo bundles an app called ThinkBook Workspace, which sits at the bottom of the longer display and lets you view widgets, a searchable clipboard history, a virtual second screen or any app of your choice. Considering that there’s an expandable screen and a motor to move it, the weight of 3.73 pounds isn’t too bad. The biggest downside to this laptop appears to be its $3,499 expected price. But you’re paying for innovation.

Read More: Lenovo is bringing its rollable, expandable-screen laptop to market – ThinkBook Plus Gen 6’s display gets taller with the press of a button or the wave of a hand

—Avram Piltch 

Best Workstation: Nvidia Project Digits

Nvidia Project Digits

Nvidia bills its new Digit (technically DIGITS) as a personal AI supercomputer, delivering up to 1 petaflops of AI computing (FP4) from its new GB10 Blackwell GPU in an incredibly small package. The system comes with a jointly developed Nvidia Arm processor designed in collaboration with MediaTek, and it could herald the first steps of companies, both of whom can sell systems based on the CPU, taking a broader step into both the desktop and laptop PC markets.

The Grace CPU under the hood, a staple in Nvidia’s Arm-based datacenter chips, wields 20 cores and communicates with the GB10 Blackwell GPU over an NVLink chip-to-chip (C2C) interface to provide efficient and fast data transfer while the unified coherent memory space reduces maximizes the memory space. Speaking of memory, Nvidia crammed 128GB of LPDDR5x memory and a 4TB SSD into the chassis. That gives the developer target market enough legroom to run up to a 200-billion-parameter AI model. For those who want to run larger models, Nvidia has also integrated its ConnectX networking, allowing you to combine two units to run up to 405-billion-parameter models.

Perhaps the most impressive facet is the form factor. Nvidia has packed an incredible amount of compute into a tiny NUC-like form factor that can easily be set on a desk — you could even imagine it bolted to the back of a monitor with a VESA mount. Naturally, such a small chassis raises concerns about thermals, but Nvidia doesn’t seem too worried — the tiny system only has two tiny air inlets on the front of the stylized front panel. You can leverage Linux and a robust suite of CUDA tools to get up and running developing AI models that can later be ported to much larger systems either in the cloud or on-prem, but the dev work can be done in this one surprisingly compact box.

Yes, performance density such as this comes at a cost — the Digit personal supercomputer costs $3,000. For now, Digit only supports Linux, but it’s possible that support for Windows could be added in the future. That would open up a broader range of use cases for the underlying CPU IP, which both Nvidia and MediaTek could use to create other derivatives, like desktp and laptop PC chips, if they choose.

Read More: Nvidia's Project Digits desktop AI supercomputer fits in the palm of your hand — $3,000 to bring 1 PFLOPS of performance home

—Paul Alcorn 

Best Gaming Handheld: Lenovo Legion Go S

Lenovo Legion Go S — Powered by Steam OS

The Steam Deck revolutionized the world of PC gaming when it launched in 2022. Since that time, we’ve seen a number of companies, including Asus, MSI and Lenovo introduce their own handheld PC game consoles. However, the problem with these Steam Deck competitors is that they all run Windows 11, which doesn’t provide an optimal UI for navigating with a pair of joysticks and some buttons.

Lenovo’s new Legion Go S — Powered by Steam OS (that’s the full name) is the first handheld not made by Valve to feature Valve’s gamer-friendly flavor of Linux. Steam OS provides an ideal entrance into Valve’s Steam platform and game store and now you can get it on Lenovo hardware.

The Legion Go S, which is also available in a Windows flavor, weighs just 1.63 pounds and is powered by an AMD Ryzen Z2 Go or Ryzen Z1 extreme processor. It features an 8-inch, 1920 x 1200 display that boasts 500 nits of brightness and 97 percent coverage of the DCI-P3 color gamut. 

The device comes with up to 32GB of LPDDR5X RAM and up to 1TB of SSD storage. However, if you want more storage, you can replace the built-in SSD with a 2280 M.2 SSD of your choice after you buy. The Legion Go S — powered by Steam OS comes in an attractive Nebula Violet (Purple) color and starts at just $499.

Read More: Lenovo’s Legion Go S handheld is powered by AMD’s Z2 Go, runs SteamOS or Windows

—Avram Piltch

Best SSD: Phison's E28 controller

Phison's E28 controller

Phison was first to market with widely available PCIe 5.0 consumer SSDs. They remain the fastest drives around, but they have one critical drawback: They draw a lot of power and can run quite hot, basically requiring active cooling. That's largely because the E26 controller uses a 12nm fabrication node. Phison will address this shortcoming with its second generation PCIe 5.0 controller, the E28, which is fabricated using TSMC's N6 process node.

The E28 has eight channels and 32 Chip Enables (CEs), with support for capacities up to 32TB. It also supports NAND speeds up to 4200 MT/s. Thanks to the process node shrink and architectural updates, power use ends up at 8.5W under heavy load — the E26 drives could hit up to 12.3W in our testing. It also has sequential read and write speeds of 14.5 GB/s, effectively maxing out the PCIe 5.0 x4 interface, with random read/write IOPS of 3,000K.

Retail SSDs using the Phison E28 controller should become available later this year.

—Jarred Walton

Best RAM: Corsair DDR5 Custom Lab

Corsair DDR5 Custom Lab

Corsair has been offering up its peripherals in custom-decorated fashion through its Custom Lab site for a couple years now, but now the company is expanding the service to its Vengeance memory, offering eight designs to start, across its three heat spreader colors: black, white, and gray. Speeds at launch will range from 6000-6,400MT/s, and CAS latencies are said to be between 30 and 36. The company expects to offer more designs soon. 

Sure, not everyone needs a tiny mural on their memory kit. But in a world where streaming and social sharing are key to many people's PC-building experience, we're often looking for new ways to customize our setups. And frankly, it's just nice to see some aesthetically pleasing RAM that isn't dominated by familiar company logos. 

—Matt Safford

Best Cooler: Cooler Master Atmos II

Cooler Master Atmos II 

Cooler Master’s original Atmos AIO impressed us with its performance, and now the company is back with a followup which adds quite a bit of customization and functionality to the mix. The block above the CPU is slimmer, and the company will offer a few swappable tops that attach via magnets. You can choose from simple caps in different colors, a circular LCD screen, or a hexagonal pixel display. There’s even a cover that incorporates a small fan for VRM cooling.

The company says it will offer the cooler with multiple radiator sizes, and sell the swappable top plates either as a kit or individually. If this cooler can maintain or improve the level of performance we saw from its predecessor while delivering all these aesthetic options, it should be one of the more appealing AIOS of 2025.

Read More: Cooler Master reveals V-Series air coolers resembling high-performance engine blocks alongside new cooling solutions at CES

—Matt Safford

Best PC Case: Cooler Master MasterFrame 400 / 600

Cooler Master MasterFrame 400 / 600

Cooler Master’s new MasterFrame breaks the PC down to its absolute basics before letting you build it back up in the specific form you choose. Based around an aluminum frame (in the 400,  and larger 600 models), the company plans to offer interchangeable panels in more forms than we’ve seen before. You’ll see the familiar tempered glass, mesh, and aluminum, and the company showcased wood, frosted glass, textured acrylic, and stone panels as well.

The company still seems somewhat unsure how it will sell additional panels and which materials will come to market, but we were told that pricing will start below $200 for the MasterFrame 400, and around $200 for the MasterFrame 600. Even in its basic vented aluminum panel form, the MasterFrame looks great. We hope, though, that many panel options will be offered at reasonable prices, because it would be fun to swap out panels to give your PC a drastically different look with minimum effort. I’d certainly keep a case longer if it could change with my tastes, or even my mood.

Read More: Cooler Master showcases three new cases and a pre-built at CES 2025 — the cases cover a wide range of aesthetics, form factors, and sizes for every user

—Matt Safford

Best Fan: Thermaltake TS120 / 140

Thermaltake TS120 / 140 

When most PCs decide how much to run their fans, they are making that decision based on the temperature of the CPU or GPU or other components, but what if the fans are on the other side of the case where it’s much cooler or hotter? Thermaltake’s new TS120 and TS140 fan series have built-in temperature sensors they can use to analyze conditions wherever you place them — the top, bottom, front or side of the case.

Using the company’s upcoming TT Plus RGB 3.0 software, these fans will be able to adjust their speed based on the region of the case they live in. You’ll also be able to see exactly how warm that area of your PC is.

The TS120 and TS140, available in three packs of block or white for $89 or $94, also have attractive ARGB lighting on both the fan rings and sides. They aren’t the the very first case fans to have temperature sensors — Corsair came out with one in 2023 — but they make the technology affordable and open to anyone who wants fine control over their thermals.

Read More: Thermaltake's new case fan has built-in temperature sensor to help adjust its speed

 —Avram Piltch

Best Gaming Monitor: Asus ROG Swift OLED PG27UCDM

Asus ROG Swift OLED PG27UCDM 

Asus’s ROG Swift OLED PG27UCDM is one of a few 27-inch monitors that debuted at CES with both a 4K resolution, OLED and a 240 Hz refresh rate. The monitor is among the first with this resolution and also boasts 99 percent coverage of the DCI-P3 color gamut, peak brightness of 1,000 nits and a response time of 0.03 ms.

MSI and Gigabyte both also had monitors with the same specs and likely the same exact panel. However, the Swift OLED PG27UCDM stands out by using a proximity sensor to tell when you walk away from the screen and put it to sleep. With OLED monitors, burn-in is a serious concern so turning them off when you’re not around is a valuable feature that will help keep  your investment working for years to come.

Read More: MSI and Asus announce 27-inch 4K QD-OLED gaming monitors with 240Hz refresh rates ahead of CES

 —Avram Piltch 

Best Accessory: Corsair Xeneon Edge

Corsair Xeneon Edge

It’s always great to have an extra screen or two for more desktop real estate. Regular monitors are great for multitasking and being productive, but what if you just want some glanceable information like your schedule, a stock price or a list of your latest Slack messages?

Corsair’s upcoming Xeneon Edge monitor is a 14.5-inch, 2560 x 720 touchscreen landscape display that’s made to show widgets like a calendar, the weather or a list of shortcuts to open your favorite apps. The monitor connects via HDMI or USB-C alternate mode and has a series of magnets that allow it to connect to an included stand or even to the outside of your case.

You’ll also be able to mount the Xeneon Edge in a special mounting panel if you buy one of Corsair’s cases. And you will be able to use the company’s iCUE software to have it show your computer’s CPU usage, temperature and other key stats.

Read More: This tiny Corsair touchscreen could change the way you view info on your PC

—Avram PIltch

Best 3D Printer: Creality Hi Combo

Creality Hi Combo

Creality’s flagship K1 printer was already excellent, providing high-quality output at speeds as high as 600 mm/s. However, the K1 isn’t perfect and there are a few things that Creality fans had wanted from it, including the ability to sustain higher temperatures and work with more exotic filaments.

Enter the Creality K1C, which takes everything we like about the K1 and builds upon it. The new printer features an all-metal hotend which allows it to operate at a steady 300 degrees and therefore work with carbon fiber (that’s what the C in K1C is for). The printer also has a revised bowden tube path, better build plate and other improvements we don’t know about yet (the company has not released full specs). One of the leading 3D printer companies has a new leading device.

Read more: Creality unveils new flagship K1C 3D printer and latest Ender 3 V3 model

—Avram Piltch

Best DIY Accessory: Hyte Powered USB Hub

Hyte Powered USB Hub

If you've ever built a system with a plethora of RGB, cooling, or other internal devices that need power and data to communicate so they can be controlled within software, you've probably run into a situation where you run out of USB 2.0 headers to plug things into. One solution is to buy an inexpensive splitter or header expander, but these aren't pretty and can be hard to hide. 

Hyte's Powered USB hub is the luxury solution to your USB header woes, turning one header into four, while encasing the PCB in an attractive translucent package with a magnetic back so you can easily place it wherever you want on the steel frame of your case. And because all the ports are on one side, cable routing should be easier and cleaner It also has a built-in light that helps illuminate your internal wiring in a pinch, and a circular knob lets you easily turn the hub off. It might  feature overkill for many internal USB needs, but it's not super expensive at $27.99, and it may be the last internal hub you'll ever need. 

Read More: Hyte shows off new PC cooling toys, including new Thicc Q80 AIO

—Matt Safford

Nvidia caught the imaginations of many technology enthusiasts on Tuesday with its show-closing, 'just one more thing'- style reveal of the Project Digits AI supercomputer. But seasoned graphics chip designer Raja Koduri and disruptive AI server startup Tiny Corp were not enchanted by the small golden box or blinded by Jensen's shimmering jacket. Both have publicly scorned Nvidia's Grace Blackwell AI Supercomputer.

At first glance, the purported provision of 1 PFLOPS of performance in such a tiny device might seem like a vision of the future, even if it is priced at $3,000. As told by the Nvidia CEO, we might believe that Project Digits is akin to AI alchemy in a box. But all that glistens isn’t gold. The spec of the headlining GB10 Superchip, the “1 PFLOP FP4 AI Compute,” is a somewhat deceptive hint to Raja Koduri and Tiny Corp (founded by American security hacker George Hotz, AKA geohot).

Koduri’s Project Digits takedown is what you probably could describe as a gentle geeky joke with a sprinkling of cynicism. The graphics chip design icon posted, “Divide flops by 4 and multiply dollars by 2. A CES (20+25)² tip for staying grounded,” poking fun at the green team.

Koduri later elaborated that - in contrast to the big FP4 claims - by his calculations, the FP16 performance of the Project Digits AI supercomputer wasn’t that impressive. Koduri estimated that the FP16 performance of the upcoming GeForce RTX 5070 and even the $250 Intel Arc B580 “seems close” to what a Project Digits machine could muster.

You may remember last year's news coverage involving Tiny Corp and its TinyBox project. In brief, Tiny Corp and its founder were very publicly displeased with AMD’s drivers. However, its plan to democratize AI acceleration by using (AMD) gaming GPUs instead of even more costly pro/enterprise alternatives was perhaps too bold. The firm managed to grab the attention of AMD publicly, and even Dr. Lisa Su and the red team were quite gracious despite the startup’s abrasive posturing and demands.

An AI swindle?

Fast forward to CES 2025, and Tiny Corp hasn’t lost its abrasive edge. “People are begging to get swindled by a $3,000 box that says AI on the side. We always get asked if we'll launch something at that price point. We won't,” grumbled Tiny Corp before advising its casual followers to “Just buy a gaming PC.”

Earlier in the day, Tiny Corp had become seemingly irritated by the Project Digits announcement. “This is marketing. FP4 is unusable, it's 500 TFLOPS of FP8. tinybox green has 4 PFLOPS of FP8, 8x more powerful.”

Nvidia’s Project Digits won’t be available until May when it is expected to roll out for around $3,000. TinyBox computers, starting from $15,000, are available for the ‘red’ model with 738 FP16 TFLOPS and “mediocre” driver quality (another barb cast at AMD).

It is difficult to overstate the performance of the best OLED gaming monitors running at 240 Hz. Not only are you getting a stunning picture with saturated color and deep contrast, but they also deliver perfect motion resolution with smooth play action and low input lag.

Though this category remains at the high end of the price scale, it is possible to buy a 27-inch QHD OLED panel for less than $600 as long as it comes from AOC’s Agon Pro line. Last year, I reviewed the, AG276QZD and now, I have its successor, the AG276QZD2, for evaluation. It adds a Quantum Dot layer for more color, and in my tests, it proved slightly quicker. Let’s take a look.

AOC Agon Pro AG276QZD2 Specs

With most 27-inch 240 Hz OLEDs costing north of $800, the AG276QZD2 represents a price breakthrough. It isn’t a revolution, but we’re making progress. Resolution is QHD, 2560x1440 pixels, which works out to 109ppi, an ideal density for sharp images and high frame rates. In the OLED world, you’ll want to hit at least 150 fps to eliminate motion blur. Above 200 fps, there’s no blur, and QHD means you can achieve this with a wider variety of video cards.

The AG276QZD2’s main upgrade over the AG276QZD is the addition of Quantum Dot color. Indeed, the gamut is now 14% larger than before. That’s a difference you can easily see in a side-by-side comparison. It also means more impact for HDR content and SDR if you choose to run the native gamut all the time. The AG276QZD2 promises and delivers accuracy with an individual factory calibration for each unit. I confirmed the included data sheet in my tests. Calibration is not required, though I achieved a small improvement with a few adjustments to the color temperature. You also get an accurate sRGB mode should you need it for photo and graphics work.

Gaming is everything I’ve come to expect from a speedy OLED monitor. At the full 240 fps, motion is incredibly smooth, and there is no perceptible input lag. The AG276QZD2 carries FreeSync and G-Sync certifications from AMD and Nvidia. There is no additional strobe feature, but it is certainly not needed here.

So, is there a “but” here? After all, there’s no free lunch; something must be sacrificed at the altar of value. Yes, there is one thing missing, and that’s high brightness. The AG276QZD2 operates with constant brightness, meaning the peak output is lower. It’s rated for DisplayHDR 400, but that level only occurs in small highlights. And SDR brightness peaks at a measured 251 nits. This doesn’t affect contrast; you still get the deep blacks that all OLEDs are known for. But if you need a really bright monitor, this isn’t it. I’ll talk more about that in my test notes later.

The AG276QZD2 doesn’t skimp on other features. You get a pair of internal speakers, USB ports, LED lighting, game aids and solid build quality with slick styling. Ultimately, the value quotient is very high and once you start playing on it, you’ll be hooked.

Assembly and Accessories

A colorfully adorned clamshell box opens to reveal the AG276QZD2’s three main components packed in molded pulp rather than crumbly foam. They mate with a captive bolt and a snap to form a solid package that exudes the high quality one expects from AOC’s Agon Pro line. A separate box houses the external power supply, and you get USB, HDMI and DisplayPort cables. A stamped steel adapter with fasteners is provided for use with arms that conform to the 100mm VESA mount standard.

Product 360

The Agon Pro line of monitors are identified by a signature shape, which the AG276QZD2 uses in the base, OSD button and panel back. It’s an angular form where each side is a different length. You won’t find anything like it anywhere else and it’s pretty cool. The all-metal base is thin with a dark silver matte finish that carries to the upright and trim. The attachment point is ringed by LED lights that glow indirectly in various colors and effects. The back is faceted with an Agon logo offset in the same dark silver.

The stand features a handle at the top and a cable management clip at the bottom. It has full ergonomics with 5/25 degrees tilt, 28 degrees swivel, 5.1 inches (130mm) height and a 90-degree portrait mode. Movements have an appropriately premium feel that is smooth and sure.

The AG276QZD2’s side profile takes advantage of OLED thinness to create a dramatic taper up to the top edge. The screen has similar reflection management to other shiny OLEDs, so you’ll need to take a bit of care when placing it. The bezel is narrow and flush with a 21mm trim strip across the bottom and an LED that glows white when the power’s on and orange in standby.

The input panel is centered up and under and features two each of DisplayPort 1.4 and HDMI 2.0. USB is version 3.2 and includes one upstream and two downstream ports. The internal speakers play cleanly with five watts of power each and you get a 3.5mm headphone jack with volume control in the OSD.

OSD Features

The AG276QZD2’s OSD appears when you press the hexagonal joystick around the back right. I found a small nit to pick with this shape because it makes the clicks less positive if you press the stick too casually. Occasionally, I would either miss a control input or errantly do a double-click.

AOC sticks with a menu design unique to the Agon Pro line with a vertical format that announces “Agon” at the top. It is ringed by red graphics and divided into eight sections.

Game Setting has six game modes in addition to the seven Eco picture modes. This can be confusing since there are so many possible combinations. My advice is to leave Game Mode off and Eco set to Standard. It allows for a full calibration and offers a choice of color gamuts with three gamma presets and three color temps. You can also engage RGB sliders for a precise result that is free of visual errors. HDR content adds four more modes, and you can use one of three HDR emulations with SDR content.

Game aids include two aiming points, one of which is dynamic in that it changes between black and red to remain in contrast. There’s also a sniper mode, frame counter and shadow control, improving dark scene visibility.

To view two sources at once, you can engage PIP or PBP with full control of window size and position. You can also swap sides and select which source plays the audio stream.

In the color menu is a low blue mode for reading plus a gamut selector (DCI-P3, sRGB or Native) and the color temp presets. The RGB sliders start at center range so it’s easy to dial the grayscale tracking.

Light FX refers to the LED ring in the back and can be programmed for different colors and effects. You can control the light level or shut it all off if you like.

The AG276QZD2 has several panel care options that include a pixel orbiter, refresh routine, corner logo dimming, boundary dimming and taskbar dimming. I left these off for testing, but when they are engaged, they only have a minor effect on the image.

AOC Agon Pro AG276QZD2 Calibration Settings

The AG276QZD2 can be enjoyed without calibration at its default settings. I recommend leaving Game Mode off because it doesn’t improve image quality. The Standard Eco mode meets industry specs for a D65 white point, DCI-P3 color gamut and 2.2 gamma. I eked out a small improvement in grayscale tracking using the RGB sliders with the User color temp option. My SDR settings are below.

In HDR mode, there are four options available. Two of them, Movie and Picture, have considerable edge enhancement, which detracts from the already sharp image. Display is the most accurate choice. 1000 Max is not any brighter and it significantly crushes shadow detail.

Gaming and Hands-on

Later, you’ll read my comments about the AG276QZD2’s luminance and HDR color saturation, but during my gaming sessions and hands-on experiences, I had nothing but love. It delivers everything you buy an OLED for, perfect blacks, bright color, incredible depth and unmatched contrast.

Gaming is also a revelation unless you’ve already played on a fast OLED. In that case, it will be every bit as good as you remember. The AG276QZD2’s QHD resolution is a great balance point for frame rates, and I was able to maintain 240fps in Doom Eternal with detail levels maxed. The picture is super sharp and loaded with fine textures and highlights. It doesn’t have quite the punch of the brighter screens I’ve reviewed like the ViewSonic XG272-2K, but it is no slouch either. I would keep it away from sunny windows but otherwise, it’s gorgeous to look at.

Video processing is on par with every 240 Hz OLED I’ve reviewed. There is no motion blur at all, and input lag simply isn’t a factor. Precise aiming and movement are easier to achieve compared to LCDs running at the same frame rates. I verified the flawless operation of both FreeSync and G-Sync with no frame tears to report.

The wide color gamut enhanced workday tasks with deep reds, rich greens and brilliant blues. Working on photos is always fun on an OLED because color is so vibrant. It’s like viewing a fine photo print with textures and elements that pop off the screen to create a 3D feel. With pixel density at 109ppi, there are no jagged lines or rough font outlines to contend with. Small elements are sharply and smoothly rendered. The AG276QZD2’s size, shape and configuration are ideal for anything one does with a computer from work to play.

Takeaway: The AG276QZD2 delivers gaming performance consistent with all the 240 Hz and faster OLED monitors I’ve reviewed. It is on par with much more expensive screens, so it leads the pack for value. Though I observed less light output than the competition, it didn’t make my gaming or work sessions any less enjoyable. And the larger color gamut is worth an upgrade from last year’s AG276QZD.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

To compare the AG276QZD2’s performance, I’ve collected five other 27-inch QHD OLED monitors. All run at 240 Hz except for the Alienware AW2725DF, which hits 360 Hz. The others are last year’s AOC AG276QZD, ViewSonic’s XG272-2K, Pixio’s PX277 OLED MAX and Acer’s X27U.

Pixel Response and Input Lag

Click here to read up on our pixel response and input lag testing procedures.

The six monitors are incredibly close in performance except for the AG276QZD’s response result, which is 2ms behind the others. Though visual differences are small, it has a slight disadvantage in smoothness. The newer screens are free of motion blur at 240fps. And there is no measurable performance benefit to the 360 Hz Alienware. In the lag test, it hits the same 19ms as the Pixio and Acer displays. The AG276QZD2 is only 2ms slower, which is a difference that only a highly skilled player will notice.

Test Takeaway: The AG276QZD2 is fully qualified for competition with very low input lag and flawless video processing. It delivers blur-free motion over 200fps and is extremely smooth over 150fps. Adaptive-Sync works perfectly too. You can expect consistent results from any of the OLEDs here, and there is no advantage to running at 360 Hz.

Viewing Angles

The AG276QZD2 looks almost identical at all horizontal viewing angles from on-center to 45 degrees. There is a barely visible red tint, but you’ll never be able to see that in content. Gamma and brightness do not change at all. The top view is also a tad warm with a slight loss of output and lowered gamma. The AG276QZD2 is completely sharable by two users.

Screen Uniformity

To learn how we measure screen uniformity, click here.

I’ve observed excellent screen uniformity in almost all the OLEDs I’ve reviewed and the AG276QZD2 is one of the best. With a 3.83% deviation, you won’t see anything but perfectly smooth tones from edge to edge. I saw this in all color patterns and at all brightness levels. There are no flaws here.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

To read about our monitor tests in-depth, please check out Display Testing Explained: How We Test PC Monitors. We cover brightness and contrast testing on page two.

Uncalibrated – Maximum Backlight Level

AOC claims 250 nits peak brightness for the AG276QZD2, which I measured in my test. The issue here is that there is no variable brightness option like most OLEDs have. This means you can’t get the brighter highlights of that feature. It’s not a big deal in SDR mode unless you really need the 300 or 400 nits offered by the other monitors. But for HDR content, you’ll miss that extra impact that comes from small extra bright bits like reflections from rendered surfaces. The good part is that you still get OLED’s infinite contrast and perfect black levels.

After Calibration to 200 nits

Calibration doesn’t change anything except peak brightness. 200 nits is a good level for indoor use in a space with moderate ambient light. When comparing SDR content, all the screens here have the same depth and contrast, whether static or intra-image. HDR is a different story which I’ll tell on page five.

Test Takeaway: The AG276QZD2 has the same SDR contrast and black levels as any other OLED. Without variable brightness, you can’t run up to 400 nits peak, but that isn’t a problem for most environments. And since it has Quantum Dot tech, there’s a good deal more color saturation to provide image impact.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

The AG276QZD2’s game and picture modes can be confusing as there are many possible combinations. To simplify, I recommend leaving Game Mode off and Eco Mode on Standard, the factory defaults. Calibration is not required, but a few tweaks net a small improvement in grayscale tracking.

Grayscale and Gamma Tracking

Our grayscale and gamma tests use Calman calibration software from Portrait Displays. We describe our grayscale and gamma tests in detail here.

Out of the box, the AG276QZD2 is a tad warm in the brightest steps with gamma that rides just below the reference. The values indicate that it’s too light, but since contrast is so high, there’s no problem. With a few tweaks of the RGB sliders, I got all errors below 1dE while gamma remained consistent. The sRGB mode removes the user color temp option but is accurate enough not to require any changes. Overall, this is excellent performance.

Comparisons

OLED gaming monitors are generally accurate out of the box. The AG276QZD2 doesn’t need calibration like all others except the Pixio which definitely benefits from adjustment. Once they’re all tweaked, the differences are invisible, and the numbers are impressive. AOC wins this contest with a 0.45dE result.

The AG276QZD2 performs well in the gamma test too with a very tight 0.03 range of values and a tiny 1.82% deviation from 2.2. The actual value was 2.16. Though this is a little light, it doesn’t negatively impact content. It truly doesn’t get much better than this.

Color Gamut Accuracy

Our color gamut and volume testing use Portrait Displays’ Calman software. For details on our color gamut testing and volume calculations, click here.

The AG276QZD2’s color gamut is significantly larger than the AG276QZD thanks to Quantum Dot tech. You can see in the default chart that every primary goes past the DCI-P3 spec by just a little. This delivers more visible color without straying too far from the standard targets. Overall error levels are as good as any professional monitor can boast. Calibration makes a subtle difference by tightening up secondary hues. The sRGB gamut is also very good with just a bit of bonus red. This is excellent performance.

Comparisons

None of the screens have visible color errors, but the AG276QZD2 runs close to the others with a 1.50dE score. As you can see, it doesn’t get much better than that. Even without calibration, the value is 1.96dE so either way, you’re seeing very accurate color.

In the gamut volume test, the AG276QZD2 takes the win with over 110% coverage of DCI-P3. That’s tremendous saturation that can only be equaled by a few pro monitors. In a gaming monitor like this, it adds a lot of texture and vibrance to the image. And if you’re in need of sRGB reference capability, the 103.74% score is nearly ideal.

Test Takeaway: The AG276QZD2 has very accurate grayscale, gamma and color right out of the box that can be slightly improved with calibration. Color volume is among the highest available from any monitor. From a value standpoint, it doesn’t get much better. It is one of the very best in its class.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

Our HDR benchmarking uses Portrait Displays’ Calman software. To learn about our HDR testing, see our breakdown of how we test PC monitors.

The AG276QZD2 offers five HDR modes, three of which can also be used with SDR content. The extra two are Display and 1000 Max and have their own unique characteristics that I’ll explain in a moment. Once you select a mode, the switch between SDR and HDR is automatic.

Of the five HDR modes, I strongly recommend avoiding Movie and Picture. They add a lot of unattractive edge enhancement that ultimately reduces sharpness. Game is free of artifacts, but it has a very light luminance curve that dulls the image. 1000 Max suggests a higher peak level, but it is no brighter than the others. It also has very dark shadows that make it unsuitable for gaming. Display is the default and best choice for its accuracy and correct tone mapping.

HDR Brightness and Contrast

It’s hard to ignore the AG276QZD2’s relatively low peak output level in the HDR brightness test. This is because there is no variable brightness feature. That omission is shared by the AG276QZD, but it is visibly brighter in HDR mode, as are the other screens. Black levels and contrast are the same across the board, so you still get the deep image OLED is known for. But the highlight areas don’t quite have the pop that they should.

Grayscale, EOTF and Color

In the HDR grayscale test, there are no errors of consequence, just a slight warmth in the brighter steps which does not adversely affect content. The EOTF tracks a bit dark, but shadow detail remains visible.

Color rendering is similar in all five HDR modes, so I’m showing the results from Display. Saturation is on target except for the 100% points, which are a bit under the mark. Most content is unaffected, but images with a lot of bright primary colors will have a little less verve. The BT.2020 chart is much the same with coverage that extends to 85% red, 75% green and 95% blue.

Test Takeaway: Though the AG276QZD2’s color gamut is wider than last year’s AG276QZD, its HDR rendering doesn’t take full advantage of the extra saturation. It definitely looks more colorful in SDR mode but HDR comparisons between the two screens are more alike than they should be. Other than overall saturation, the AG276QZD2 shows excellent color accuracy and solid tone mapping.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

When it comes to OLED gaming monitors, it is still true that you won’t find a bad one. They are super consistent in performance and appearance with tremendous contrast, color saturation, accuracy and build quality. That said, I’ll always laud a monitor that offers good value.

The AOC Agon Pro AG276QZD2 is that monitor. It has the same contrast as any OLED with its unmeasurable black levels. It has a huge color gamut thanks to Quantum Dot tech which is larger than last year’s AG276QZD. Accuracy comes from a factory calibration that can only be slightly improved with adjustment. It also delivers the best gaming performance available from any monitor with perfect motion resolution and low input lag.

My only complaints are that it is less bright than the competition and that the most vivid HDR colors are a bit muted. The brightness issue comes down to its lack of variable brightness. Most of today’s OLEDs have this feature, but AOC has left it out. That means SDR and HDR peaks are just over 250 nits which is lower than the competition. But so too is the cost. At less than $600 at this writing, there isn’t a better deal in the OLED category. And in practice, the HDR color issue was minor. For the price, I can forgive.

If you’ve been dying to put an OLED gaming monitor on your desk but balked at the entry fee, the AOC Agon Pro AG276QZD2 is worth serious consideration. It delivers everything OLEDs are known for including that addictive gaming experience, which makes it worthy of your consideration.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

Intel announced that its first wave of updates to address the gaming performance issues plaguing the Arrow Lake processors is now available. Intel says the updates available today will deliver "signficant performance upside" and restore the "complete and intended functionality" of the Core Ultra 200S Arrow Lake processors. We're currently testing the fixes and have summarized the root causes below.

Intel's update comes after a disappointing round of Core Ultra 200S launch day reviews, which found wildly variable gaming performance recorded by different media outlets. Those results overwhelmingly did not match the performance benchmarks that Intel had shared before the launch, leading to an overall poor reception for the Arrow Lake processors. Last month, Intel's Robert Hallock announced that an investigation was underway, saying the company planned to restore gaming performance to its original targets and would share details this month. 

The update requires both a new motherboard BIOS firmware and a Windows update that also corrects several issues. Intel advises its customers to update to the latest BIOS from their motherboard vendor and to update Windows to build 26100.2314 (or newer). 

In January, Intel will issue a final component — a new BIOS revision that provides additional performance beyond what was originally promised. These are features and improvements developed after that launch. Intel says the final update will add an additional single-digit percentage increase in gaming performance. At CES 2025, Intel will issue a full A/B performance analysis of the issues and new benchmarks reflecting the performance of the entire fix.

Intel identified five root causes, four of which are already addressed by updates available today. We'll cover Intel's list of the root causes, and then we'll give a synopsis of the fix for each.

Intel Arrow Lake Root Causes

• 1. A missing Performance & Power Management (PPM) package. 
• 2. Intel Application Performance Optimizer (APO) could not take effect. 
• 3. BSODs when attempting to launch game titles utilizing the Easy Anti-Cheat service. 
• 4. Select performance settings misconfigured on reviewer or early enabling BIOSes. 
• 5. New BIOS performance optimizations.

1.) The Performance & Power Management (PPM) package dictates how the Windows Power Plans control the CPU's behavior. The PPM is typically delivered through a chipset driver of Windows Update. Intel says that a missing or malfunctioning PPM impacts multiple performance facets (details in the accompanying blog post), thus reducing performance by 6 to 30% (est.), depending on the workload or other extenuating factors. 

Intel says it incorrectly scheduled this update for 'user/retail availability' but not for reviewer availability. The PPM is now added in Windows 11 build 26100.2161 (KB5044384) or newer. It has been available in Windows Update since November. 

During a YouTube live stream at Hot Hardware, Robert Hallock confirmed that the correct PPM package was not available to press for reviews, either in Windows Update or the chipset drivers. Intel says the impact of this can vary tremendously, so we'll have to test to see how it impacts our gaming results. 

2.) Intel's Application Performance Optimizer (APO) boosts game performance in several game titles. This software utility is now automatically installed by default in Windows, but it also had to be manually enabled in some BIOSes (we confirmed the feature was enabled in our testing). Intel says the incorrect PPM (covered above) resulted in APO not working as intended, costing anywhere from 2% to 14% of performance. This is now resolved with the addition of the correct PPM package. 

3.) The BSOD issues caused by anti-cheat software have been resolved by an Easy Anti-Cheat driver update that Epic Games is distributing to the relevant publishers and developers. 

4.) Intel cites that the motherboard BIOSes provided to reviewers had incorrect default settings for a range of "VIP settings." The number of incorrect settings and the settings impacted varied by vendor. These included settings like "PCIe Resizable BAR, Intel APO, compute tile ring frequency, memory controller ratio (gear), sustained/transient power limits, and more." 

Intel cited a whole host of potential effects, including "aberrantly high memory latency (~1.5-2.0x expected); erratic or fluctuating compute tile ring frequency; high run-to-run stdev [standard deviation] for dynamic or unpredictable workloads; no performance uplift for games that benefit from PCIe Resizable BAR; no performance uplift for games that benefit from Intel APO." 

Intel says this impacted performance anywhere from 2% to 14% (the same as the performance loss from APO not working correctly). Intel said it did not correctly enforce the VIP settings with motherboard vendors. All new BIOS revisions correct this issue. 

5.) Intel will issue a new BIOS and CSME firmware kit in the first half of January 2025, but the availability will vary depending on the vendor. This final update will include an undisclosed "small selection of performance optimizations" developed after the launch. Intel says this update will add another single-digit percentage of gaming performance (as measured by Intel with a geomean of ~35 games). 

This new firmware update is Intel microcode version 0x114 and Intel CSME Firmware Kit 19.0.0.1854v2.2 (or newer). The 0x114 microcode will not work correctly unless paired with the CSME Firmware Kit. The CSME is fully encapsulated within the BIOS update, so it doesn't require other driver updates. 

Thoughts

Intel says the currently available Windows and firmware updates will bring Arrow Lake to the promised performance levels — the newer BIOS coming in January is just an additional level of performance beyond Intel's original projections. We've also seen that software updates, such as the Cyberpunk 2077 update that increases performance by up to 33%, can also significantly improve Arrow Lake's standing. Both of those are promising developments. 

We've also seen beta testing of the new  0x114 microcode slated for release in January. Those initial tests show little to no performance improvements, but that is with the caveat that the firmware requires the accompanying CSME Firmware Kit to work correctly, which isn't available yet. That invalidates the testing. We'll have to wait for the official versions of the 0x114 microcode to pass judgment. 

In the meantime, all that's left to do is test Intel's claims that the performance issues are corrected with the updates available today. As you can imagine, we're already busy testing the changes. Stay tuned.   

The best gaming monitors don’t have to be pretty to be functional. Top performance is certainly a key element, but gaming is much more fun when your display looks the part. Many manufacturers add an attractive aesthetic to their gaming screens, but AOC’s Agon line goes up a level from the rest thanks to a relationship with the famed Porsche Design.

Porsche Design isn’t only known for penning legendary sports cars with engines in the back. It also lends its name and expertise to various products, from watches to sunglasses. And it was indeed founded by F.A. Porsche, who conceived the original 911. If that isn’t a timeless piece of rolling art, I don’t know what is.

AOC has produced several gaming monitors with Porsche Design’s help, and the latest example is the Agon Pro PD49. It brings high performance and high style to a 49-inch 32:9 QD-OLED panel with 240 Hz, Adaptive-Sync, HDR400 and wide gamut color. Let’s take a look.

AOC Agon Pro PD49 Specs

Curved monitors have gone from the answer to a question no one asked to an essential gamer’s tool in just a few years. The pinnacle of the format is the 49-inch 32:9 category. This layout is becoming more popular as an alternative to putting two 27-inch screens on the desktop. With DQHD 5120x1440 resolution, you can have 109ppi pixel density and lots of screen area with no dividing line.

The PD49 brings every bell and whistle to the party. It’s an OLED with an 1800R curve, which is enough for greatness on its own. To that, it adds a Quantum Dot layer for wide gamut color. And it’s very wide; indeed, I measured over 115% coverage of DCI-P3 in my tests. That’s a whole lotta color. There is bonus saturation in every primary. Out-of-box accuracy is decent as well. Calibration isn’t strictly required, but there has been some improvement after a few tweaks.

The PD49 is certified for HDR 400 TrueBlack, and it easily met this spec in my tests using a 25% white window. With constant brightness turned off, you can expect peaks around 1,000 nits from small highlights. And as a premium OLED, it delivers the same perfect black levels as every other OLED I’ve reviewed. The picture quality is stunning.

With a 240 Hz maximum refresh rate, smooth motion is no problem. There is no need for overdrive, and there is no strobing option. But it isn’t too difficult to run at 240fps. And at speeds over 150fps, there is no motion blur whatsoever. Measured input lag is also very lo,w so you won’t have a problem dominating opponents in your favorite virtual battles.

Physically, the PD49 is nothing less than industrial art. The stand is a substantial piece of cast aluminum that weighs more than some complete monitors. The back of the panel is adorned with vertical lines and logos that glow with soft LED lighting. There are plenty of inputs, all the latest versions, DisplayPort 1.4, HDMI 2.1 and a USB-C for video. KVM is supported by one upstream and four downstream USBs. A pair of eight-watt internal speakers deliver quality audio with DTS tuning and user tweakability.

The PD49 sells for $1,200 at this writing, which is on par with other monitors in its class. The small upcharge for elegant design is appropriate. You can save a few bucks if you just want a competent 49-inch OLED gaming monitor but none I’ve seen look as good when the power is off.

Assembly and Accessories

The PD49 comes in a coffin-like enclosure with its contents protected by flexible foam rather than the crumbly stuff. The stand is already assembled and bolts onto the back of the panel with four included fasteners. A nice Phillips-head screwdriver is included for that purpose. If you’d rather use an arm, a 100mm VESA mount is provided. The cable bundle includes IEC power, two USBs, DisplayPort and HDMI. A USB thumb drive, by Porsche Design of course, contains manuals, drivers and an app to make the OSD accessible from the Windows desktop.

Product 360

It’s hard for today’s thin bezel monitors to set themselves apart from the front view, but the PD49 has a stand that can only be described as monolithic. It’s made from six heavy aluminum castings fused into a super solid mass that is a pleasure to look at. It offers full ergonomics with 3/13 degrees tilt, 20 degrees swivel and a 5.1-inch height adjustment. Movements are firm and smooth with the feel of a finely tuned machine. The PD49 is incredibly overbuilt and should endure many years of use.

The screen sports a thin flush bezel that’s a half-inch wide around the top and sides and one inch at the bottom. The front logo says Porsche Design, while AOC and Agon get their visibility in the back, where those logos include LED backlighting. That can be set to flash, breathe or remain steady in a variety of colors. Or you can turn them off for a stealth look. Another LED adds a soft glow to the desktop. The vertical line theme carries over with shiny molded in bits set off against a matte background. The stand includes a small cable management clip.

The 1800R curvature strikes a good balance between immersion and practicality. There are curvier screens out there, but the PD49 avoids any possibility of image distortion. It is extremely useful for productivity as well as entertainment. It is just like having two 27-inch QHD monitors, except there’s no dividing line.

The input panel is well stocked with one DisplayPort 1.4, two HDMI 2.1 and a USB-C for video sources. HDMI supports 2560x1440 resolution at 120 Hz for consoles. PCs can enjoy the full 5120x1440 at 240 Hz through DisplayPort or USB-C. There is also a KVM hub with one upstream and four downstream USB ports. A 3.5mm jack supports headphones and there are two internal speakers driven by eight watts of clean power. DTS tuning raises the audio quality to a high level, and you can tweak eq settings in the OSD.

OSD Features

All Agon Pro monitors, like the PD49, get AOC’s cascading menu rather than the bottom of the screen ribbon. A small joystick in the center of the panel’s bottom edge controls everything. Or you can use the app included on the thumb drive to change settings from the Windows desktop.

The OSD is divided into eight sections, starting from the bottom with Game Setting. It has a selection of Game Modes, shadow control and boost for lighter blacks and a color slider to increase saturation. Game aids include a sniper mode and two aiming points that can be toggled by clicking the joystick towards you. One is always red and the other changes between red and black to maintain contrast with the background. You can also engage a frame counter.

The Luminance menu has picture modes (Standard is the only one you’ll need), three gamma presets and the Uniform Brightness option. Turn this off to see the PD49’s maximum light output. I did all my tests and gaming with it off. When it’s on, the picture is darker in tone and somewhat dull unless you turn all the room lights off and play at night.

The HDR options are found in Image Setup, where four additional HDR picture modes provide subtly different luminance curves to suit different content. Hint: stick with HDR TrueBlack for the best picture in all situations. HDR emulation is possible in SDR mode with the other three HDR modes. This menu also has PIP and PBP options to view two video sources at once.

In Color Setup, you can choose between three fixed color temps and a user mode with RGB controls. There are 10 gamut options that cover all standards currently in use from sRGB to BT.2020. Panel Native is the best choice there because it uses the PD49’s full capability which covers over 115% of DCI-P3.

The Audio menu offers five DTS sound modes plus TruVolume HD which enhances sound dynamically. If you want to tweak, there’s also a five-band equalizer. Light FX is the LED lighting feature, and it can be set to a variety of effects and colors. It employs the two backlit logos on the back plus a soft pool of light for the desktop.

The final menu, Extra, has several panel care options including a pixel orbiter and refresh tool. You can also set an off timer, change the image aspect ratio and return all settings to their factory defaults.

AOC Agon Pro PD49 Calibration Settings

The PD49 measured well out of the box in my tests and doesn’t strictly need calibration, but I found the image to be a tad warm in tone. A few tweaks of the RGB sliders took care of that and tightened up color saturation tracking as well. The improvement is subtle but worth the effort. My recommended settings are below. For HDR content, stick with TrueBlack as it is the most accurate choice with correct luminance tracking and near-perfect color. I also recommend leaving Uniform Brightness off for maximum image impact.

Gaming and Hands-on

Though OLED gaming monitors represent a premium (expensive) category, one thing about them is undeniable, all examples at a particular refresh rate perform the same. Where LCDs can differ in the quality of their overdrives and strobe features, OLEDs will always provide perfect motion resolution above 200fps. The PD49 showed no motion blur, no frame tears, and no artifacts at any time during my gameplay sessions. And its input lag is among the lowest of any monitor. A professional player might want the super low lag of a 500Hz screen but for most of us, the 21ms delivered by the PD49 will be plenty quick.

In my experience, this means precise aim and movement. I always play Doom Eternal because it’s familiar so comparisons are easy and accurate. Every 240 Hz OLED I’ve reviewed gives me the same feel. Run and gun maneuvers are easy because I can turn and shoot without overcompensating, even when the action is frantic. The aim point stops precisely on target, so I don’t waste ammo. And difficult jumps are nearly always successful. There’s nothing worse than erasing game progress with a fall off a cliff. The PD49 keeps that from happening.

The picture is a stunner as well. HDR is an addictive experience both for its bright highlights and rich color. The PD49 lets you play SDR games with that same wide color gamut but it’s more effective in HDR titles because that’s the way they’re encoded. Black levels are incredibly deep but never lacking in shadow detail. Though the PD49 includes two shadow enhancer controls, they are completely unnecessary because low-end gamma is on the mark. I noted gamma issues in the brighter areas though where a few highlight areas seemed disproportionately bright. This is a small problem that is more obvious to someone who plays on a lot of different gaming monitors, but it is worth noting.

For daily tasks, the PD49’s 1800R curve is perfect for keeping multiple documents and windows open. If you are a fan of multiple screens, a 49-inch 32:9 monitor delivers that same functionality without the dividing line. And at current prices, it’s less expensive than buying two 27-inch QHD OLEDs.

Things unique to the PD49 are, of course, its high style and build quality. I am not blessed with the means to own a Porsche automobile, but this monitor delivers the same feel as a luxury sports car. It’s gorgeous to look at and it drips with premium style. Moving the panel through its adjustments is an uplevel experience from most of the monitors I encounter.

I also noted the PD49’s excellent sound quality. Large monitors make it possible to include better speakers and AOC has delivered here. While there isn’t a ton of bass, the sound is rich and full, with excellent clarity. At half volume, I more than filled my personal space with game music and effects. The DTS modes are useful too with distinct differences heard between them.

Takeaway: The PD49 performs on par with the other 49-inch 240 Hz QD-OLED monitors I’ve tested. It’s extremely colorful and accurate. It delivers the contrast all OLEDs are known for. It’s the perfect gamer’s tool with no motion blur and super quick response. And it’s equally well suited for work and play. Plus, it rises above the rest with premium styling and build quality.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

I filled out the PD49’s comparison group with 49 (and one 45) -inch OLEDs running at either 144 or 240 Hz. They are Philips’ 49M2C8900, Samsung’s OLED G9, AOC’s AG456UCZD, Aorus’ CO49DQ and Asus’ PG49WCD.

Pixel Response and Input Lag

Click here to read up on our pixel response and input lag testing procedures.

240 Hz typically means a 4ms draw time for a full white field pattern. The PD49 meets this standard. Like all OLEDs, it has no motion blur above 150fps. There is no strobe option here, but none is needed. I saw no flickering or artifacts whatsoever.

The PD49 takes the input lag contest by 1ms over the other AOC and Philips displays. If control response is your priority, 240 Hz should be a minimum requirement. The two 144 Hz screens won’t hold up a casual gamer like me, but if your skills are high, so should your frame rates. It’s worth the price premium to have a monitor and PC that can keep the action over 200fps.

Test Takeaway: The PD49 is fully qualified for competition with super low input lag and perfect motion resolution. Though most gamers won’t be toting a 49-inch monitor to LAN parties, it is the perfect display for those who battle from their basements. To get the most from it, I recommend a well-equipped PC that can keep the fun at 200fps or higher.

Viewing Angles

The PD49 shows a slight red shift when you move to 45 degrees off-center. This is typical of the QD-OLEDs I’ve photographed. The extra screen layer that delivers more color also has a slight polarizing effect. But there is no loss of light output, nor is gamma changed. The PD49 is completely shareable. The top view also retains its brightness and gamma with no apparent change in color.

Screen Uniformity

To learn how we measure screen uniformity, click here.

With a 1.86% deviation from the center zone, you’d think the PD49 would have set a record for my test database which as of this review, contains 407 monitors. But that title still belongs to the Asus PA32UCG I tested in 2021 which scored 1.05%. No matter, this is still an incredible result. There is no doubt that the PD49 is all about quality, inside and out.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

To read about our monitor tests in-depth, please check out Display Testing Explained: How We Test PC Monitors. We cover brightness and contrast testing on page two.

Uncalibrated – Maximum Backlight Level

OLED tech hasn’t come up to the brightness level of Mini LED yet, but the PD49 and its competition deliver more than enough output for any indoor space. With any OLED, I always recommend leaving Uniform Brightness off so the panel will modulate brightness dynamically. It allocates power to the brighter elements and ultimately gives the image more pop. For this reason, I measure a 25% window pattern for all OLED reviews. The PD49 peaks at the same 449 nits as the PG49WCD, excellent performance. Black levels are too dark to be measured by currently available instruments so contrast cannot be determined.

After Calibration to 200 nits

Adjusting the 25% window to 200 nits lowers the peak brightness just enough to be comfortable in a dim-to-dark room scenario. Black levels and contrast are unchanged. The ANSI test shows the same results. I can’t measure the black squares of the checkerboard pattern so intra-image contrast is theoretically infinite.

Test Takeaway: OLED isn’t the brightest tech out there but with its perfect black levels, it provides the most realistic and textured image possible in a display. The PD49 is on par with all OLEDs in all size and resolution categories. It looks stunning, but no more so than any other OLED.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

The PD49 can be enjoyed right out of the box without adjustment. But I found the picture a tad warm in tone when using the default Standard picture mode. You can tweak to taste or calibrate the user color temp as I did.

Grayscale and Gamma Tracking

Our grayscale and gamma tests use Calman calibration software from Portrait Displays. We describe our grayscale and gamma tests in detail here.

The first chart shows the PD49’s slight red errors that can be seen in areas above 70% brightness. They are small enough to be forgivable and won’t significantly impact most content. The gamma tracking chart shows a dip at 90% which means some bright bits will be too bright. In most cases, this isn’t a big deal, but some content might show a little less clarity of detail in highlight areas.

After adjusting the precise RGB controls, there is no visible grayscale error, but gamma is unchanged. The other presets are further off the mark, so I stuck with Gamma 1. This isn’t a huge issue, but I would prefer to see closer adherence to the 2.2 reference.

If you choose the sRGB gamut option, there is no visible effect on grayscale which is a good thing. The color temp options are grayed out in this mode. This chart shows that the PD49 is qualified for color grading or photo editing.

Comparisons

The PD49 has grayscale performance similar to that of its competition. All the screens here are good, and there are no significant issues to report. In the gamma tests, the PD49 is close to the others in average value, but its range of values is a little larger than the rest thanks to the dip at 90% brightness which cannot be adjusted away. This is a minor issue that is far from a deal-breaker.

Color Gamut Accuracy

Our color gamut and volume testing use Portrait Displays’ Calman software. For details on our color gamut testing and volume calculations, click here.

The PD49 excels in all its color gamut measurements. I checked out Panel Native and sRGB which closely match DCI-P3 and sRGB standards, respectively. The other modes also adhere to industry specs. You can have Adobe RGB if you want, which is something rarely found outside the professional realm.

The initial run shows a hue error in magenta and slight over-saturation of all colors. If you want to get closer to DCI-P3, that option is available and delivers an average error of just 1.82dE. However, a few users will complain about more color. Panel Native was my choice for gaming and general use. Calibration reduces saturation slightly and brings magenta closer to its targets.

sRGB is spot-on apart from the red primary, which is a tad over-saturated. However, with an average error of 1.75dE, it is close enough for critical applications.

Comparisons

OLED monitors constitute a tough crowd when it comes to color accuracy comparisons. The PD49 finished fifth, but with a 2.29dE score, it is free of any visual errors and more than accurate enough for its intended purpose. The big win comes in the gamut volume test where it cracks the 115% mark. It is the most colorful OLED monitor I’ve tested to date. And with 105% coverage of sRGB, it is nearly ideal when that gamut is needed.

Test Takeaway: The PD49 isn’t quite perfect out of the box, but it’s close enough that calibration isn’t required. But with a few tweaks, the picture hits the reference level for grayscale and nearly so for gamut accuracy. A slight gamma anomaly keeps it from attaining perfection, but this is a minor complaint. The PD49 delivers a gorgeous image dripping with saturated color to the naked eye.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

Our HDR benchmarking uses Portrait Displays’ Calman software. To learn about our HDR testing, see our breakdown of how we test PC monitors.

The PD49 accepts HDR10 signals and switches modes to reveal four additional presets. TrueBlack is the default and best choice because it’s color-accurate and follows the correct luminance curve so that detail remains clear in all parts of the image. The other modes have slightly different looks and can be used as emulators with SDR content if you wish.

HDR Brightness and Contrast

In my travels thus far, only the AG456UCZD has exceeded the 600-nit barrier for ultra-wide monitors. And it’s 21:9 versus the others’ 32:9 so it’s something of an outlier. All the 49-inch models look the same in a side-by-side comparison with nearly the same output levels. The only difference will be in color saturation, where the QD models stand apart. The PD49 is one of those so its HDR image is a sight to behold. I measured the other modes and found no differences in peak output or contrast. There is no brightness penalty for sticking with TrueBlack.

Grayscale, EOTF and Color

The PD49’s HDR grayscale tracking has the same slight warm tone that I saw in the SDR tests. This is a minor issue that will be hard to spot in content. There is no calibration option, but that’s not a problem. The EOTF chart shows tight tracking with slightly dark levels from zero to 40%, then a tad above the line until the tone-map transition at 65%. This is excellent performance.

In the color tests, the PD49 is a bit over-saturated across the board which is typical of the HDR monitors I’ve tested. This gives the image more impact and sets it further apart from SDR content. Detail rendering remains sharp and clear with no loss of detail. In the BT.2020 test, the PD49 covers 90% red, 80% green and 95% blue. This is more than any non-QD monitor can boast. You won’t find more color anywhere.

Test Takeaway: The PD49 delivers a stunning HDR image with bright highlights, sharp detail and beautifully saturated color. Accuracy is high and there are no visible issues to report. If you want the best possible HDR, get an OLED like this and you’ll never go back to LCD.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

In the world of large gaming monitors, the 49-inch 32:9 category is the most extreme. It’s ideal for first-person gaming, whether you play shooters or racing sims. And if you can afford an OLED, the experience is truly up-level.

For those able to spend a little more money, there’s AOC’s Agon Pro Porsche Design PD49. It offers the same superb video processing and stunning image of other QD-OLEDs, but it adds high style and unsurpassed build quality for a little more money.

Performance-wise, the PD49 is on par with its competition. Motion resolution at high frame rates is entirely free of blur. Input lag is very low, 21ms in my test. This makes it the quickest 49-inch 240 Hz OLED I’ve tested. Gaming on it is an addictive experience.

The image is stunning, like every OLED, and further enhanced by Quantum Dot tech. The PD49 is currently the most colorful OLED I’ve reviewed, with over 115% coverage of DCI-P3. Though I noted a slight gamma anomaly, it can be enjoyed without calibration. The picture is very accurate, and my adjustments made only a small improvement.

Sound quality is definitely above average. With eight-watt internal speakers, excellent clarity and volume, and DTS tuning with multiple modes, you can enjoy an immersive experience without headphones.

The PD49 is a premium gaming monitor in every respect. From its superb build quality and styling to its high performance and stunning image, it is a purchase that no one will regret. And since it isn’t priced significantly higher than its competition, you’re essentially getting that Porsche Design cachet for only a small upcharge. At this writing, it’s selling for around $1,200, so if that is within your budget, the PD49 should be on your short list.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

The Crucial P310 2280 takes everything we liked about the original Crucial P310 (2230), then stretches it to fit the M.2 2280 form factor. The original 2230’s combination of surprisingly good performance and excellent power efficiency can now be had in a longer package for your desktops and laptops. But there's tons of competition in the 2280 market, so how does the new variant stack up? Budget SSDs continue to surprise us, and the P310 looks to be no different — but there are a lot of rivals in this space.

Yes, you could buy the 2230 version and it will fit most longer M.2 slots — possibly with an adapter for certain laptops. This might even make sense for M.2 2242 hosts, but in general you pay a premium for 2230 drives. This makes a 2280 version of the P310 a pretty good idea, especially from Crucial’s viewpoint. It uses the same hardware, so the drive can remain single-sided and cool-running with essentially no changes. That’s good, as the P310 had good performance for a QLC-based drive, with very high power efficiency. It also means that maybe the drive should be more affordable.

If you’re looking for a budget drive that doesn’t feel like a budget drive, you could do worse than the P310 2280. Its everyday performance feels similar to that of a high-end drive. It does have the dreaded QLC flash that some dislike, but in real world terms and especially with how Crucial has designed this drive, it's often not a big deal. On the other hand, a drive built on QLC flash should be less expensive than its TLC-based peers even if the drives feel similar 99% of the time. The P310’s primary drawback is that with the added competition found in the M.2 2280 form factor, Crucial has to take pricing more seriously.

The Crucial P310 2280, as in the M.2 2230 form factor, comes in 1TB and 2TB capacities. The 2280 variant also offers up a 500GB SKU, which is interesting from a budget perspective but probably more for OEMs that just want a modern SSD as cheap as possible. This makes more sense for the 2280 as the 2230 is more typically used to upgrade existing internal drives that already start at 512GB or more.

Current pricing, now well over a month after the original NDA, comes in $5 below the specified MSRP, but we're in the middle of the holiday shopping madness and prices have fluctuated a lot. The MSRPs for the P310 2280 were far too high and would have made the P310 a pretty bad deal, with targets of $64.99, $94.99, and $149.99 respectively for the 500GB, 1TB, and 2TB SKUs. Current street prices are sitting at $41.99, $71.99, and $144.99.

That's a good sized and necessary price drop of 35% and 25% on the two smaller models, while the largest model, and the one we're reviewing, only sees a current 3.3% discount. However, inventory at some retailers has dried up and we did see pricing as low as $129.99 (13% off) earlier in the month. Considering the competition, the P310 2280 2TB really should be priced closer to $119.99, perhaps even lower, as plenty of good quality TLC drives land in that price range.

In general, we recommend TLC-based drives over QLC models, particularly when pricing is close. TLC is even more important for the low capacity 500GB drive, as that's only 125GB of raw NAND capacity (i.e. pSLC cache) and it's very easy to fill up such a drive so that there's little room for the cache. Drives like the Teamgroup MP44L or WD Blue SN580 are probably a better bet at 500GB. These have a lower performance ceiling, but with so few flash dies you are better off with TLC.

The 1TB price has come down a lot since the initial launch, but reasonable alternatives like the Kingston NV3 and Teamgroup MP44 still undercut the P310 2280. The same goes for 2TB, with the SN580 sitting at $99.99, and drives like the Addlink A93, Silicon Power US75, and MSI M482/M482 Eco-Pack can fall below $100 as well. Crucial has, historically, priced its drives competitively when needed, so we expect prices to come down more on this.

The P310 2280 is not short on performance, at least. It can reach up to 7,100 / 6,000 MB/s for sequential reads and writes and up to 1,000K / 1,200K IOPS random read and write IOPS. This is a ton of performance for a budget drive, although we are talking QLC flash here. Crucial backs the drive with a five-year warranty with up to 220TB of writes per TB capacity.

Crucial P310 Software and Accessories

If you travel to Crucial’s website for P310 downloads you’ll be greeted with a wide array of information including written and video-based guides. This is a nice touch. At the top is a driver for Crucial and Micron NVMe SSDs, although this is unnecessary — the stock Windows driver is fine. Crucial also offers its Crucial Storage Executive, which is an SSD toolbox with all the typical features you’d expect: drive health information, the ability to perform drives wipes and firmware updates, and more.

Also available is Acronis True Image for Crucial, which will assist you in migrating data to a new drive with backup and imaging functionality. The full suite does help Crucial stand apart from much of the competition, aside from perhaps Samsung, although expert users will likely have their own solutions. Our drive also came with one month of the all apps plan for Adobe Creative Cloud, which doesn’t mean much aside from the fact Crucial is targeting this drive at a range of users — Adobe likely paid to offer that preview deal would be our guess.

Crucial P310: A Closer Look

The 2TB P310 is single-sided and, quite frankly, bereft of anything interesting. That’s a good thing as less is more when it comes to SSDs. Since we’ve already reviewed it in a smaller form factor where it's single-sided with just one NAND package, that’s all we would expect here. We also know the design is DRAM-less, so there should only be a controller and no DRAM.

With everything revealed, we find that we are correct. This is the same package found on the 2TB M.2 2230 version of the drive, although it's oriented differently and further away from the controller. This makes use of the extra space to potentially improve thermals by spreading out the components a little. It’s still using sixteen 232-Layer QLC flash dies in a single package.

One possible observation here is that, if the hardware is the same, you’d expect the 2230 and 2280 versions of the drives to have the same price. In fact, the M.2 2230 market segment has less competition especially at 2TB, so Crucial can make the P310 more expensive. That makes sense, and it means that buying the 2230 version to extend it isn’t necessarily a good proposition outside of maybe for a 2242 device.

Conversely, you might think you can buy the 2280 version for less money and cut it down — a real practice that many people have engaged in for other drives. However, Crucial’s placement of the NAND flash package on the 2280 version is such that this is impossible. That might be another reason they put more space between it and the controller.

Interestingly, the P310 2230 2TB drive actually costs $139.99 right now, $5 less than the 2280 model. The price on that drive was slashed by $40 on November 21, just in time for the Black Friday leadup. So, at this particular snapshot in time, the smaller model could be a viable option. Just be sure to check the latest prices before buying, as they tend to fluctuate quite a bit.

MORE: Best SSDs

MORE: Best External SSDs

MORE: Best SSD for the Steam Deck

Comparison Products

The P310 has stiff competition, which only helps demonstrate how far budget drives have come. The fastest PCIe 4.0 drives are old now, but they're still performant even if they're also less power efficient. These include the SK hynix Platinum P41, the Samsung 990 Pro, and the WD Black SN850X. When bringing things down to four channels for better efficiency, we have Crucial’s T500 with DRAM, the new Samsung 990 EVO Plus without DRAM, and the Silicon Power US75. The last one shares hardware with many popular, “high-end” budget SSDs on the market with 232-Layer TLC flash but no DRAM.

Samsung’s 990 EVO is a step further down, filling in for the slower budget drives. In the opposing direction, we have the Phison E31T ES, the basis of budget PCIe 5.0 drives that will improve upon existing designs. The true high-end includes drives like the Crucial T705 and the Adata Legend 970 Pro that are in a class of their own, but cost more as a result.

Trace Testing — 3DMark Storage Benchmark

Built for gamers, 3DMark’s Storage Benchmark focuses on real-world gaming performance. Each round in this benchmark stresses storage based on gaming activities including loading games, saving progress, installing game files, and recording gameplay video streams. Future gaming benchmarks will be DirectStorage-inclusive and we also include notes about which drives may be future-proofed.

The P310 2280 lands right in the middle of the pack. This is impressive as it beats drives like the SN850X, but it's not able to keep up with some other drives. Considering it’s DRAM-less, this is a good result. Keep in mind that performance on a fuller drive can drop as there's less spare area to work with.

Trace Testing — PCMark 10 Storage Benchmark

PCMark 10 is a trace-based benchmark that uses a wide-ranging set of real-world traces from popular applications and everyday tasks to measure the performance of storage devices. The results are particularly useful when analyzing drives for their use as primary/boot storage devices and in work environments.

The P310 2280’s showing in PCMark 10 is even better. Here it’s only behind the T500, which has DRAM and TLC flash of the same generation. Crucial doesn’t want to cannibalize its own products but that shouldn’t be a problem as anyone who wants DRAM and/or TLC flash will go with the T500. Both the T500 and P310 are excellent choices for laptop users with the T500 leaning enthusiast, but PCMark shows that for many workloads they perform similarly.

Console Testing — PlayStation 5 Transfers

The PlayStation 5 is capable of taking one additional PCIe 4.0 or faster SSD for extra game storage. While any 4.0 drive will technically work, Sony recommends drives that can deliver at least 5,500 MB/s of sequential read bandwidth for optimal performance. In our testing, PCIe 5.0 SSDs don’t bring much to the table and generally shouldn’t be used in the PS5, especially as they may require additional cooling. Check our Best PS5 SSDs article for more information.

Our testing utilizes the PS5’s internal storage test and manual read/write tests with over 192GB of data both from and to the internal storage. Throttling is prevented where possible to see how each drive operates under ideal conditions. While game load times should not deviate much from drive to drive, our results can indicate which drives may be more responsive in long-term use.

The P310 2280 could be a good choice for a PS5 drive, if price comes down. If you want 4TB you'll have to look elsewhere, and the P310 500GB probably isn’t a good idea either — you could fill that with two or three larger games these days. It makes the most sense at 2TB and there’s a lot of competition there. At that point, it often comes down to whatever drive is least expensive at the moment. Assuming you at least want something at the upper end of PCIe 4.0 performance, right now you would be better off with the Klevv CRAS C925, MSI Spatium M482, and some others.

Transfer Rates — DiskBench

We use the DiskBench storage benchmarking tool to test file transfer performance with a custom, 50GB dataset. We write 31,227 files of various types, such as pictures, PDFs, and videos to the test drive, then make a copy of that data to a new folder, and follow up with a reading test of a newly-written 6.5GB zip file. This is a real world type workload that fits into the cache of most drives.

The P310 2280 does well in DiskBench, with copy performance second among the tested PCIe 4.0 drives, only falling behind the T500. Both drives push the boundaries of 4.0 with same-generation flash, but QLC flash has a lower write performance ceiling. Performance here is good enough — within the large pSLC cache, remember — to compete even with higher-end drives like the SN850X and 990 Pro.

Synthetic Testing — ATTO / CrystalDiskMark

ATTO and CrystalDiskMark (CDM) are free and easy-to-use storage benchmarking tools that SSD vendors commonly use to assign performance specifications to their products. Both of these tools give us insight into how each device handles different file sizes and at different queue depths for both sequential and random workloads.

The P310 2280’s ATTO curves are smooth with good performance at all block sizes. Sequential performance in CDM is not as rosy, as the drive’s QLC flash limits what it can do against the TLC-based competition. We think its performance is still good enough. For a drive of this class, where it’s not really supposed to be doing workloads like sustained writes, small I/O performance — indicated by 4KB, QD1 random latencies — is more important.

The P310 2280 has exceptionally low write latency, and even if this is within the pSLC cache it reflects real-world feel as by far and large your random writes should be cached. Random reads will come from native flash, and while QLC flash is not the fastest there are architectural improvements designed to make it more responsive to such demands. As such, the P310’s read latency is also quite good for its interface speed.

Sustained Write Performance and Cache Recovery

Official write specifications are only part of the performance picture. Most SSDs implement a write cache, which is a fast area of pseudo-SLC (single-bit) programmed flash that absorbs incoming data. Sustained write speeds can suffer tremendously once the workload spills outside of the cache and into the "native" TLC (three-bit) or QLC (four-bit) flash. Performance can suffer even more if the drive is forced to fold, which is the process of migrating data out of the cache in order to free up space for further incoming data.

We use Iometer to hammer the SSD with sequential writes for 15 minutes to measure both the size of the write cache and performance after the cache is saturated. We also monitor cache recovery via multiple idle rounds. This process shows the performance of the drive in various states as well as the steady state write performance.

The 2TB P310 writes at 6.3 GB/s for around 64 seconds in the single-bit pSLC cache mode. That makes the cache size around 400GB, which matches pretty well with our findings in the 2230 review. This is a large cache but is not the largest possible. 4-bit QLC flash in a single-bit mode could lead to a cache size of 500GB. By using a smaller cache, the P310 can try and avoid folding — the moving of data from pSLC to QLC flash, which is forced if space must be freed — for a little bit of time, or sustain a higher speed after the cache runs out.

After the cache runs out, the P310 is forced to write to QLC flash while moving data over from pSLC to QLC in order to free up space for incoming and upcoming writes. In this mode it writes at 337 MB/s on average, which is quite slow by most standards. That might seem awful, but it’s better than the 100 MB/s or so we saw with Crucial’s older P3 Plus 2TB.

The newer flash on the P310 is almost twice as fast to write with each die, but it also has 50% more planes and that can cap its interleaving potential sooner. The result is that the 337 MB/s we see here is actually not that bad. Consider that the HP FX700, which has a comparable controller and similar QLC flash from YMTC, only manages 139 MB/s. This suggests that Crucial’s trimming of the pSLC cache is effective in improving sustained performance, even if this is below maximum SATA speeds.

As for recovery, the drive mostly stayed in its slower state even with some idle time. Normally, large-cached drives will recover pSLC rapidly, although this can increase wear. Because Crucial went with a slightly more modest cache it’s able to be more patient, which probably improves the user experience in the long run. One reason for this is that it can retain data in pSLC longer, and that can improve follow-up reads in addition to reducing wear by deferring writes. On the whole, this is a pretty good strategy for a QLC-based drive.

Power Consumption and Temperature

We use the Quarch HD Programmable Power Module to gain a deeper understanding of power characteristics. Idle power consumption is an important aspect to consider, especially if you're looking for a laptop upgrade as even the best ultrabooks can have mediocre stock storage. Desktops may be more performance-oriented with less support for power-saving features, so we show the worst-case.

Some SSDs can consume watts of power at idle while better-suited ones sip just milliwatts. Average workload power consumption and max consumption are two other aspects of power consumption but performance-per-watt, or efficiency, is more important. A drive might consume more power during any given workload, but accomplishing a task faster allows the drive to drop into an idle state more quickly, ultimately saving energy.

For temperature recording we currently poll the drive’s primary composite sensor during testing with a ~22°C ambient. Our testing is rigorous enough to heat the drive to a realistic ceiling temperature.

The P310 2280 is very power-efficient, and that’s no surprise. The combination of the excellent Phison E27T controller with the newest QLC flash is bound to produce a drive that doesn’t pull a lot of power. The controller is DRAM-less with only four channels but is still able to max out the PCIe 4.0 interface, allowing the drive to finish workloads quickly. This is especially true with a workload that fits in the pSLC cache. As a result, the P310 is pretty amazing for laptops.

A lower power draw can mean better thermals, too. In our testing the P310 hit a maximum reported temperature of 52°C. This is far below any throttling threshold and this drive should be great when run without a heatsink.

We use an Alder Lake platform with most background applications such as indexing, Windows updates, and anti-virus disabled in the OS to reduce run-to-run variability. Each SSD is prefilled to 50% capacity and tested as a secondary device. Unless noted, we use active cooling for all SSDs.

Crucial P310 2280 Bottom Line

The Crucial P310 did not disappoint when we reviewed it in the M.2 2230 form factor, and it doesn’t disappoint now in the larger form factor. Performance is surprisingly good across the board when you consider the drives it’s up against. Sustained performance does remain a problem for QLC-based drives like this, but Crucial does a halfway decent job of mitigating that. Power efficiency is superb, which helps make this drive a great choice for laptops on a budget — or at least it would if the price comes down. The MSRP was way too high for a drive like this — and that’s nothing new, from Crucial or any other manufacturer — but even the street prices remain too high.

This is something it shares in common with the Samsung 990 EVO Plus, an excellent drive that we had to dock half a point due to its pricing and late arrival. Now, the 990 EVO Plus uses TLC flash and as a result has an edge over the P310. We’re merely pointing out how both brands are struggling with the pricing aspect, in part because that’s just how the industry works. You can’t dilute your brand with a low MSRP and the volatile flash and SSD markets mean pricing can be challenging, especially for a new product where inventory is still ramping up.

In any case, docking the P310 2280 half a point over its M.2 2230 sibling has more to do with the fact that the shorter form factor has a lot less competition. 2230 is specialized and the P310 stands out more in comparison to the alternatives. In fact, the 2280 P310 looks somewhat average in this review, but it’s facing some of the fastest drives out there. This is a testament to how good a budget drive can be, even with QLC flash. It would probably stand out more with a 4TB option, but it’s quite good at 2TB and likely would be at 1TB, too. We can safely recommend it for laptops, desktops, and the PS5 at those capacities. 500GB seems like an afterthought, but for an ultra-budget build it could still make sense.

Looking at 2TB and current pricing, we can safely say that the P310 2280 is the best QLC-based drive out there. It’s a known brand with new but known hardware. However, its real competition has TLC flash and this includes drives like the MSI Spatium M482, Lexar NM790 (and newer Lexar Play 2280), Inland TN470, Teamgroup MP44, and more that are at the same price or lower. This is discounting the eight-channel drives with DRAM — some of which are double-sided, and all of which will run hotter in a laptop — as well as lower-end budget PCIe 4.0 drives like the Teamgroup MP44L. If you’re going for a budget PCIe 4.0 drive right now, you might be inclined to look at the top-end 7+ GB/s drives, although on a very tight budget you can get away with the MP44L and drives in its class.

This leaves the P310 2280 in a bit of an odd spot because it feels like a high-end product for what it is, but simultaneously it's designed to be in the budget class of drives. It’s DRAM-less and has QLC flash, which more or less precludes it from being high-end, even when it sometimes performs that way. We suspect it will end up being a popular drive either way, and the price will come down. It’s just a matter of Crucial positioning it correctly, and historically they have been able to do so.

3.5 stars is a good score in our rankings, and the drive holds its own against the competition. However, we know some people will always look down on QLC flash. That’s fine, but anybody looking for a budget SSD that still has some speed — whether as a primary drive for their laptop or budget desktop build, or as a secondary drive, or even in the PS5 — can pick up the P310 2280, install it, and never have to second-guess the decision.

MORE: Best SSDs

MORE: Best External SSDs

MORE: Best SSD for the Steam Deck

Mesh routers are a popular way to expand coverage in your home, especially in areas that a traditional router might underserve. By strategically placing wireless nodes throughout your home, you can cover areas like the workshop in your garage or your outside patio to enjoy sunlight and nature as you work from home.

Now that Wi-Fi 7 proliferates through the best Wi-Fi routers, we’re seeing more mesh systems using the standard at various price points. Asus’ new ZenWiFi BQ16 Pro is a more premium offering, with a street price of just over $1,100 for a router and one satellite. However, Asus packs the ZenWiFi BQ16 Pro with features, including two 6 GHz bands, a tri-band MLO wireless backhaul, support for 10 Gbps wired backhaul, and coverage of up to 8,000 square feet.

In addition, the 6 GHz performance on the ZenWiFi BQ16 Pro is the fastest we’ve seen in a router. However, is the dazzling 6 GHz band performance enough to justify dropping over $1,000 on a mesh router? Read on to find out.

Design of the Asus ZenWiFi BQ16 Pro Wi-Fi 7 Mesh Router 

I was shocked when our review unit arrived because the box was quite large and heavy. The size and weight of the box were more akin to something like a desktop-replacement gaming laptop than a typical wireless router. Inside the box were the two wireless nodes (router and satellite), power cables for each device, and a patch cable for connecting the router to your modem.

Each node is large, measuring 8.4 x 6.9 x 2.8 inches, and finished in white plastic. The design reminds me of a large hardcover book, and neither would look out of place sitting on a bookshelf, desk, or living room end table. There is plenty of ventilation all around to keep these hardworking network devices cool, with the bulk of the cooling coming from vents on the back – if you look closely, you’ll notice that “WIFI 7” is spelled out in the rear ventilation holes, like what we saw with the Asus RT-BE86U. 

You’ll also find all of the ports on the back, including two WAN ports: one is 10 GbE, the other is 1 GbE. This dual WAN functionality provides a fallback to continuous internet connectivity (if you have two ISPs). There’s one more 10 GbE LAN port and two additional 1 GbE LAN ports. For such a high-end router, I’m surprised that Asus didn’t at least swap in 2.5 GbE ports for the 1 GbE ports. That seems like a big miss for a product that retails for over $1,000.

You'll also find a single USB 3.0 port, a power port for the barrel-style power adapter, and a power switch. Note that the router and the satellite have the exact same rear port configuration, unlike some other mesh routers with fewer rear ports for the satellites.

Asus ZenWiFi BQ16 Pro Wi-Fi 7 Mesh Router Specifications 

Setting up the ZenWiFi BQ16 Pro Wi-Fi 7 Mesh Router

Like all of Asus' routers, the ZenWiFi BQ16 Pro can be set up using a smartphone app (Asus Router) or a web interface. Asus provides the smartphone app for those who want a quick and easy route for setting up the router and don't plan on tinkering with the hundreds of possible settings. To get the most out of the router, I used the web interface for setup and interacted exclusively with the router utilizing that method.

After powering up the first wireless node, which became the router, I pointed my browser to asusrouter.com. From there, the setup program walked me through some basic procedures, like naming the four primary SSIDs, creating an IoT SSID, setting an administrator login and password, and checking for firmware updates. This initial process took about five minutes in total.

After the router finished its initialization, I powered up the satellite. The setup program then synched the satellite to the router and optimized the mesh network. Once this configuration was complete, I had invested roughly ten minutes into the whole setup process for the mesh system.

Since the router and satellite use identical hardware, each has two 10 GbE ports on the back. That means you can use a free 10 GbE port for wired backhaul, boosting overall throughput between the router and satellite to provide the best network reliability. Dual WAN is also supported, allowing you to leverage up to 10 GbE for one ISP and up to 1 GbE for the second.

If all else fails with your wired WAN access, you can configure the ZenWiFi BQ16 Pro to tether to your smartphone to leverage its cellular connection using the USB port on the back. You can enable this setting from the “USB Application” menu in the web interface. The ZenWiFi BQ16 Pro also supports USB-based cellular modems.

While a wired backhaul will give you the best performance and reliability with a mesh router setup, the ZenWiFi BQ16 Pro has the next-best thing: a robust wireless backhaul. It leverages Wi-Fi 7's MLO functionality to bolster the wireless backhaul by bonding the 2.4 GHz, 5 GHz, and 6 GHz bands into a single link.

You can also create multiple virtual networks based on your needs (i.e., Guest, IoT, Kids, etc.).

Asus ZenWiFi BQ16 Pro Wi-Fi 7 Mesh Router Software 

The ZenWiFi BQ16 Pro uses the AsusWRT 5.0 software platform, and our review unit was tested using the 3.0.0.6.102_36789 firmware. You could easily spend hours getting lost in the various menus and submenus, tweaking every last bit of performance out of the router and the satellite. Asus is among the best in the business in giving you complete control of nearly every aspect of your network.

The Network Map is your router homepage, giving you a basic overview of your network status. You can see the number of connected clients, how many satellites are connected, which wired ports are active (including the USB 3.0 port), along with RAM/CPU utilization.

Suppose you want to expand beyond the satellite that comes in the box. In that case, you can utilize the AiMesh feature to expand your mesh network with another supported router (preferably, another Asus Wi-Fi 7 router). Asus also offers its comprehensive AiProtection platform, which is provided by TrendMicro. It offers features like malicious site blocking, automatic quarantine for infected devices, an intrusion protection system to protect against DDoS attacks, and an automatic scanning feature that can pinpoint vulnerabilities in your network that need to be addressed. Unlike some other vendors that charge you a monthly or yearly subscription, this functionality is provided for free.

I spent at least an hour roaming through the web interface, checking out all the features, tweaking settings to see how they would affect performance, and testing out some of the more targeted features. For example, the ZenWiFi BQ16 Pro supports Apple Time Machine backups for Macs, and it can be enabled from the USB Application window.

Enabling the feature was as simple as plugging in a USB hard drive (which I had already formatted using APFS) and turning on the Enable Time Machine button. Performing a Time Machine Backup over the network takes considerably longer than if you were using a USB hard drive directly connected to your Mac. Still, wireless backups, especially with a laptop, are a nice feature to have.

Asus ZenWiFi BQ16 Pro Wi-Fi 7 Mesh Router Performance 

As usual, our Wi-Fi router testbed relies on a Windows 11 desktop with an MSI Pro B650M-A Wi-Fi motherboard, AMD Ryzen 5 7600, 32GB of DDR5, a 1TB PCIe 4.0 SSD, and an MSI Herald-BE Wi-Fi 7 PCIe adapter.

The iPerf3 throughput tests are conducted with our Windows 11 server featuring an onboard 10 Gbps wired network card connected to the 10 Gbps LAN port on the ZenWiFi BQ16 Pro. Wireless tests encompass 6-foot and 25-foot distances.

All wireless tests are initially run with the network free of added traffic from other connected clients. Tests are then performed to simulate traffic from additional users accessing the network (for our testing, we use six clients streaming 4K video from YouTube).

Let me preface the results by giving away a major spoiler. The ZenWiFi BQ16 Pro is the fastest Wi-Fi 7 router I’ve tested to date, and it’s not even close. Starting with the iPerf3 6-foot uncongested tests on the 6 GHz band, the ZenWiFi BQ16 Pro was clocked at an astonishing 3,523 Mbps, more than doubling the result of the Netgear Orbi 770 and nearly doubling the performance of the TP-Link Deco BE65 Pro. At 25 feet, the ZenWiFi BQ16 Pro dropped to 1,907 Mbps, but that was still far ahead of the 1,106 Mbps and 845 Mbps performance of the Orbi 770 and Deco BE65 Pro, respectively.

5 GHz band performance was still good, but it wasn’t a complete blowout like with the 6 GHz tests. The ZenWiFi BQ16 Pro managed 1,630 Mbps at close range compared to 1,530 Mbps for the Deco B65 Pro. The router underperformed slightly at long distances, with 611 Mbps compared to 652 Mbps for the Deco BE65 Pro.

The tables turned when it came to the 2.4 GHz band, with the ZenWiFi BQ16 Pro trailing the other two competitors. It scored the lowest numbers in the group, delivering 111 Mbps at 6 feet and 58 Mbps at 25 feet.

The ZenWiFi BQ16 Pro was also no slouch with congested traffic, as it nearly hit 3 Gbps at 6 feet on our 6 GHz iPerf3 tests while moving out to 25 feet gave us 1,613 Mbps. Both numbers were more than double what the competitors could muster. It was a similar result with the 5 GHz band, with the ZenWiFi BQ16 Pro hitting 1,557 Mbps at 6 feet and 569 Mbps at 25 feet.

The ZenWiFi BQ16 Pro’s misfortunes with the uncongested 2.4 GHz results were flipped with congested traffic. This time, the mesh router barely eked out a win at 6 feet 108 Mbps) but came in last place at 25 feet (45 Mbps).

For most of our ping tests (congested and uncongested), the ZenWiFi BQ16 Pro stayed below the 10 ms mark on the 6 GHz and 5 GHz bands. However, those times climbed to as high as 17 ms on the 2.4 GHz band with congested traffic.

As always, we must put the disclaimer that these tests are only based on the makeup/layout of my home. Results can significantly vary depending on the architecture of your home, apartment, or office.

Bottom Line

The Asus ZenWiFi BQ16 Pro is a genuine speed demon in the Wi-Fi 7 router realm. We saw performance crest the 3,500 Mbps mark on the 6 GHz band at close range, and even the long-range tests came close to 3,000 Mbps. 5 GHz band performance was also class-competitive, while the results on the 2.4 GHz band were a bit more mixed.

My gut feeling tells me that most home users will likely be relying on the 5 GHz or 6 GHz bands for their oft-used devices (i.e., smartphones, tablets, laptops, desktops, Smart TVs, etc.), while the 2.4 GHz band will primarily be used by low-speed smart home or IoT devices. Even so, the 2.4 GHz band performance is not bad, per se; it just isn’t as impressive as what we saw on the other bands.

In addition to the performance, Asus gives you great network flexibility with 10 Gbps wired backhaul for satellites, MLO wireless backhaul, two 6 GHz bands, and numerous settings to further tweak your network for performance or reliability. You even get Dual WAN support and smartphone fallback via the USB 3.0 port. Software features abound, including free network monitoring and intrusion protection.

The only thing on my wish list would be to swap out the three 1 GbE ports with 2.5 Gbps or 5 Gbps ports. After all, this is a mesh router with a street price of over $1,000 — I think that Asus could at least spring for 2.5 Gbps ports at this price point to go along with the two 10 Gbps ports already offered. But if you’re looking for an absolute Wi-Fi 7 speed champion and don’t mind the $1,100 price tag that comes with the distinction, it doesn’t get much better than the ZenWiFi BQ16 Pro. 

When I was preparing for this review of the Acer Predator X34 OLED, I thought that I had already reviewed an Acer monitor called X34 in the past. A quick search of my archives revealed another Predator X34 from 2016 and an X34P from 2018. OK, so this is the X34 OLED. But wait a minute. Acer’s catalog currently includes four X34s, two of which have OLED panels. How to choose?

My subject here is the X34 OLED with the 800R curvature. It’s the most extremely curved monitor in Acer’s lineup and in anyone else’s for that matter. An 800mm radius means the edges of the screen practically wrap around your head. You can actually hear your voice reflected into your ears when you sit close.

But that’s not all. The X34 OLED features a 34-inch 21:9 screen with WQHD 3440x1440 resolution, a 240 Hz refresh rate, Adaptive-Sync, HDR400, wide gamut color, KVM switching and the latest connectivity. And it sells for around $900 at this writing, which puts it in the sweet spot among the best gaming monitors.

Acer Predator X34 OLED Specs

While $900 isn’t trivial, the X34 OLED is less expensive than other 34-inch 21:9 OLED gaming monitors. Concerning the naming, the X34 OLED is not to be confused with the X34 QD-OLED. That more expensive display has, you guessed it, Quantum Dot technology for a larger color gamut. It also has a 175 Hz refresh rate and a gentler curve radius. The X34 OLED, on the other hand, is built for high-performance gaming and leaves nothing under the table regarding speed and response.

The refresh rate is 240 Hz, which is achieved without overclocking, and both flavors of Adaptive-Sync are supported. Two DisplayPort 1.4 inputs and a USB-C port accept 3440x1440 signals with 10-bit color processing. Two HDMI 2.1 ports are there for consoles and other video source components. The X34 OLED has a KVM switch with four USB downstream ports (version 3.2), including two right up front with a headphone jack for easy access.

As a non-QD panel, the X34 OLED doesn’t quite have the color saturation of the latest QD-OLEDs, but that is not to say it isn’t colorful. It has more gamut volume than almost any Mini LED monitor with a measured 98.78% coverage of DCI-P3. Due to its infinite OLED contrast, it delivers a stunning image and killer HDR with peak highlights over 1,000 nits. I measured a 25% window pattern at over 660 nits, which makes it brighter than nearly all of its direct competitors.

The gaming feature list is long enough to provide essential items like aiming points, a very flexible sniper mode, refresh rate counter and timers. There is no LED lighting, but that is the only thing missing. There are plenty of picture modes, calibration options and a constant brightness option. OLED panel care comes in the form of a pixel shift and a refresh program. And a pair of internal speakers deliver decent audio.

The Predator X34 OLED doesn’t quite set a new bar for value, but at around $100-200 less than other 34-inch gaming OLEDs, it delivers a lot for the money. Let’s take a look.

Assembly and Accessories

The X34 OLED comes packed in crumbly foam in three parts for a toolless assembly. The screen is well protected by a thick peel-off plastic sheet. The base is super solid and once put together, you would have to try very hard to knock it over. The cable bundle includes IEC for the internal power supply plus HDMI and two USBs. There is no DisplayPort. You also get four hefty adaptor lugs for the 200x100mm VESA mount if you’d rather use an arm.

Product 360

The X34 OLED’s front bezel is very thin and almost invisible when the power is on. It’s even flush at the bottom so there’s no logo there to take your eyes out of that gorgeous picture. Only a tiny Predator logo appears at the bottom of the stand. The base is a full 12 inches deep and almost two feet wide, but its slender legs leave plenty of room for peripherals or random desk clutter.

The upright is also slender but rock solid with full ergonomics. You get 5/15 degrees tilt, 15 degrees swivel and 4.8 inches of height. Adjustments have the premium feel I’ve expected from all Predator displays. At the top of the upright is a quarter-inch threaded mount for webcams or other peripherals. And did I mention it’s completely made from metal, not plastic around a metal core. I’ve called other monitors monolithic, but the X34 OLED earns that adjective more than most.

An 800R curvature is currently the tightest radius available. You can see in the top-down view that the sides of the screen come quite far into the user’s peripheral vision. Though image distortion is minimal, this shape imparts something of a fishbowl effect. It’s magnified by the reflection of sound back into the user’s ears. If you like to talk out loud while gaming (I do), you’ll hear your own voice very clearly. It’s a great format for first-person games and simulators. And the thin bezel means you can line up two or three monitors for a hyperrealistic driving or flying rig.

The input panel is up and under and includes two HDMI 2.0, two DisplayPort 1.4, and three USB 3.2, one upstream and two down. Under the front edge of the screen are two more USBs, a type and a type C, along with a 3.5mm headphone jack. An LED bar glows orange in standby and blue when the power is on. A tiny joystick is the only physical control. Two internal speakers play with five watts of power and deliver clean loud sound with reasonable bass and no audible distortion.

OSD Features

Pressing the X34 OLED’s joystick summons a quick menu that lets you change inputs, adjust brightness and switch picture modes. Another press opens the full OSD, which is divided into seven sections and includes three settings memories.

The Game Assistant menu seems simple, with only three options, but the aim points and sniper mode include a lot of useful functionality. The reticle comes in three shapes and can be red, white or green. Or, you can have it change color automatically with the background to maintain its visibility. You can also move it around the screen. The sniper mode has three magnifications, multiple sizes and a night vision option. And you get four countdown timers.

The next menu, Gaming, has a FreeSync/Adaptive-Sync toggle and a frame counter. Ultra-Low Latency is available when you turn FreeSync off and reduces input lag by 1ms according to my tests. If you can run consistently faster than 200fps, it’s a viable option, but with FreeSync on, input lag is still very low.

Happily, Acer doesn’t hide the X34 OLED’s constant brightness option in the OLED care menu. It’s in the Picture menu where it’s easier to find. If you want the monitor’s full brightness capability, it should be left off. When it’s turned on, the peak white level is around 260 nits. This menu also has Black Boost for enhanced shadow detail, an HDR toggle, Super Sharpness (edge enhancement, leave off) and a Low Blue Light option.

The Color menu has eight picture modes, three of which can function as settings memories. If you make any change at all, the X34 OLED reverts to the User mode. Calibration options include gamma presets and color temps plus two-point RGB sliders with gain and bias control, very nice. A selection of color spaces is also offered but I found during testing that they all measured the same. There is no usable sRGB mode, more on that later.

Hot Key Assignment lets the user pick a function for the first two slots of the quick menu. You can also specify normal or low HDMI black levels. Leave this one on normal for best results. Low clips some shadow detail and does not improve picture quality. The second screen of System options has a pixel shifter and a refresh routine for OLED panel care. Finally, you can save your settings to one of three memories that are part of the picture mode selection, very cool.

Acer Predator X34 OLED Calibration Settings

I found the X34 OLED a little green in tone out of the box, so some adjustment is required for best results. In the user color temp option, you get a two-point control with gain and bias sliders. I also measured a dark gamma close to 2.4. Changing the preset to 2.0 and tweaking the RGB gains made a significant improvement in all aspects of the image. I’ve listed my settings below. For the brightness values, I’m showing what it takes to hit the nit values with Constant Brightness off and on. The higher number is required when it’s on. The X34 OLED responds to HDR signals with an auto-switch and grays out all picture controls.

Gaming and Hands-on

Gaming on the X34 OLED is a superlative experience. I’ve become spoiled by 240 Hz OLEDs for sure, but this is one of the faster ones. I could sense its lower input lag even though it’s only a couple of milliseconds quicker in my tests. Run and gun maneuvers were super easy because I could quickly reverse my viewpoint or direction of travel. I can only imagine this monitor in more skilled hands. It is truly competition-worthy.

WQHD resolution meant I had no trouble maintaining 240fps in Doom Eternal. This is where an OLED is in its element. There is no hint of motion blur at any time. I could not make a mouse movement fast enough to cause even the slightest smear.

The image is stunning with its deep contrast and bright highlights. I could tell the X34 OLED was brighter than other 34-inch screens I’ve reviewed. Only the Asus PG34WCDM hits higher HDR peaks. Asus is over 660 nits while Asus cracks 700.

However, I did note some shadow detail issues in HDR mode. My tests showed dark EOTF tracking, which manifested in the game. Doom Eternal has an excellent tweaker for this issue that allows the user to adjust multiple parameters to make every part of the image clear. Once I did this, the image rose to another level. I would have liked the X34 OLED to track luminance correctly and out of the box, but at least the fix isn’t difficult.

I also enjoyed the tight curvature. 800R is as tight as it gets for any curved monitor and the wraparound effect is excellent. I was pulled into the game, and I enjoyed better situational awareness since everything was within my field of vision. I didn’t have to turn my head at all. I can see the potential for a killer racing sim with two or three X34 OLEDs.

In productivity apps, I enjoyed the razor-sharp picture and deep contrast. There are no shadow detail issues that need to be corrected in SDR mode. With a change to the gamma 2.0 preset, everything is clearly visible down to the tiniest font or photo element. The picture is very colorful, and I didn’t miss the Quantum Dot tech. Its omission here is not a negative.

Regarding the 800R curvature, I always noticed it but if I kept my focus at the center of the screen, it wasn’t a problem. Web browsing and document editing are a common activity that is best performed in the middle two-thirds of the X34 OLED. Spreadsheets are a bit more unusual if you view them full screen. It just requires a little getting used to. The only thing I wished for when working in Photoshop was an accurate sRGB mode. I could only color grade in the DCI-P3 gamut.

Takeaway: Aside from its missing sRGB mode, there is nothing not to like about the X34 OLED. It’s a terrific gaming monitor with super quick response and blur-free motion processing. The color and contrast are everything I’ve come to expect and enjoy from OLED displays. And it produces decent sound from its internal speakers. The extreme curve is different, but once acclimated, it’s not a problem in productivity, and it’s a definite asset for first-person gaming.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

I had no trouble finding other 34-inch 21:9 OLEDs for the X34 OLED’s comparison, but 240 Hz examples are still relatively new; I’m up to four now. We have Corsair’s 34WQHD240-C, Gigabyte’s MO34WQC2 and Asus’ PG34WCDM. At 175 Hz are Samsung’s OLED G8 and Philips 34M2C8600.

Pixel Response and Input Lag

Click here to read up on our pixel response and input lag testing procedures.

Four milliseconds is the expected draw time for any 240 Hz monitor, OLED or otherwise. Remember that an OLED is smoother at a given refresh rate than an LCD. In other words, a 240 Hz LCD will have a little motion blur, but an OLED won’t have any. The X34 OLED has perfect motion resolution in test patterns and real-world content. For my test, I left FreeSync turned on and still recorded a super-quick 20ms total lag time. Only the Corsair is 1ms faster and I doubt even the most skilled gamer will perceive that difference. The X34 OLED is fast for sure. If you choose to run without Adaptive-Sync, total lag is 20ms.

Test Takeaway: All the 240 Hz OLED gaming monitors I’ve tested are fast. They, and the X34 OLED, are in the top category for both screen response and input lag. You can have a little less lag with a premium LCD running at 500 Hz, but then you’ll give up that magical image that only an OLED can produce. At this point in time, there is nothing better for gaming.

Viewing Angles

The X34 OLED shows the same near-perfect off-axis image as every other OLED I’ve photographed. The side view shows no appreciable difference in color, brightness or gamma. This is an ideal panel for an 800R curve because it’s tough to sit such that your eyes are equidistant from all parts of the screen. An OLED won’t penalize you for that extreme curve. The top view shows a slight gamma reduction and a bit of green.

Screen Uniformity

To learn how we measure screen uniformity, click here.

I have yet to see an OLED deliver anything but visually perfect screen uniformity. As a premium display, that should be the case and the X34 OLED doesn’t disappoint. I saw no glow, bleed or hotspots from my sample using a variety of field test patterns from full black to full white and in all primary and secondary colors.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

To read about our monitor tests in-depth, please check out Display Testing Explained: How We Test PC Monitors. We cover brightness and contrast testing on page two.

Uncalibrated – Maximum Backlight Level

Like any OLED, the X34 OLED measures luminance differently depending on the setting of the Constant Brightness option. When it’s on, the peak is around 260 nits. Off pushes the max to nearly 400 nits. My personal preference is to have it off because the picture is livelier. By the numbers though, the contrast is the same. If you turn it on, you’ll have to raise the brightness slider to achieve the same max level. Black levels and contrast cannot be measured in either case.

After Calibration to 200 nits

Calibration has no effect on black levels or contrast. In the settings table on the previous page, I’ve provided values for both settings of the Constant Brightness option. ANSI contrast also cannot be measured for the X34 OLED or any other OLED monitor.

Test Takeaway: The X34 OLED performs similarly to other OLEDs in my SDR brightness and contrast tests. I award an extra point here because the Constant Brightness option is in the Picture menu rather than hidden with the OLED panel care features. As I’ve said before, this is why you buy an OLED. No other display technology can produce this much dynamic range and it makes for a stunning image.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

The X34 OLED ships in its Eco picture mode, but that means limited brightness. You can change to Standard easily enough, but making any adjustment will switch the monitor to User mode. There, you have a full set of calibration options, and you’ll need some of them for the best possible picture.

Grayscale and Gamma Tracking

Our grayscale and gamma tests use Calman calibration software from Portrait Displays. We describe our grayscale and gamma tests in detail here.

The X34 OLED’s default grayscale shows green errors from 40% brightness and higher. They are visible in test patterns and in content. Gamma also tracks on the dark side with a value closer to 2.4, even though the default preset is 2.2.

Luckily, the fixes are easy ones. A few tweaks of the RGB gains and a change to gamma 2.0 and the X34 OLED is much closer to perfection. This is pro-level performance.

The sRGB mode is accessed by choosing that gamut from the color space list. It also shows the same green tint and dark gamma. Here, there are no options for correction.

Comparisons

A 5.13dE grayscale result is out of character for the premium OLED category. The X34 OLED definitely needs some love to look its best. The dark gamma is also a negative because it makes the image a little subdued. Fortunately, its image controls are very precise and effective. By changing the gamma preset to 2.0 and tweaking the RGB gains, it now plays in the same sandbox as the others with a 0.87dE grayscale error and gamma value of 2.16. The range of values is also much tighter than before with a winning range of just 0.12, excellent performance. And this has a positive effect on color, as you’re about to see.

Color Gamut Accuracy

Our color gamut and volume testing use Portrait Displays’ Calman software. For details on our color gamut testing and volume calculations, click here.

By default, the X34 OLED’s color tracking isn’t too bad. It’s closer to the mark than grayscale, with an average error of 2.89dE. The hue error in cyan is visible in content, but the other colors look good. After calibration, the error is lower by half. Every point is on target. It doesn’t get much better than that.

Though the X34 OLED has a full set of color gamut options, they all look and measure about the same. The sRGB mode still renders DCI-P3 color as you can see in the third chart above. While a usable sRGB mode isn’t an absolute requirement for a gaming monitor, the choice should be there for those who want that smaller gamut for SDR content or color grading applications.

Comparisons

The X34 OLED’s final color result is one of the best I’ve recorded in this display category. A 1.08dE average is about as low as it gets. In the comparison, it takes second place but in practice, you won’t see a difference between any of these screens except in overall saturation. The X34 OLED isn’t a QD-OLED, so it won’t be quite as colorful as the top four monitors. But it’s a tad less expensive too, so that’s worth considering. And 98.78% coverage of DCI-P3 is pretty close to the ideal 100%. sRGB is a non-starter, with a very over-saturated 146.46% score. If you really need sRGB, the X34 OLED is not the monitor for you.

Test Takeaway: The X34 OLED delivers extremely accurate color, grayscale and gamma once calibrated. It’s out-of-box performance is a little disappointing, as is its lack of a usable sRGB mode. But with a few tweaks, it becomes a gorgeous monitor. Apply the settings I’ve recommended here, and you can enjoy the same high performance.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

Our HDR benchmarking uses Portrait Displays’ Calman software. To learn about our HDR testing, see our breakdown of how we test PC monitors.

The X34 OLED responds to HDR10 signals with an automatic switch to HDR mode if that option is turned on in the OSD. It is by default. There are no image adjustments available, nor are there additional picture modes.

HDR Brightness and Contrast

Most HDR400 monitors peak around, surprise, 400 nits. Some top 460 like the Corsair. However, the X34 OLED and the Asus take it further with 660 and 703 nits, respectively, when measuring a 25% window. In practice, this means brighter highlights and a more vibrant picture. This is a difference you can plainly see. OLEDs aren’t as bright as their Mini LED counterparts, but the X34 OLED is among the brightest in its category. Even though it doesn’t have Quantum Dot color, its extra punch could be a purchase decider. Black levels and contrast are unmeasurable.

Grayscale, EOTF and Color

The X34 OLED’s HDR grayscale test shows the same slight green tint that I saw in the SDR measurements. It’s visible between 40 and 100% brightness. Its impact on actual content varies with what you’re looking at. The EOTF tracking starts out too dark and gets a little light before hitting the tone-map transition at 70% brightness. It might be necessary to compensate in-game to see all the available shadow detail.

The X34 OLED does well in the HDR gamut test with solid tracking of the DCI-P3 standard. Red and blue are slightly over-saturated but no more so than any other HDR monitor I’ve tested. It’s become standard to push primary colors in HDR displays to add some punch and make a more visible difference between SDR and HDR content. It isn’t strictly correct, but it is forgivable because it has no negative impact. The X34 OLED isn’t a Quantum Dot panel so i,t can’t fill as much of BT.2020, but it makes a good run with accurate tracking. I have no complaints here.

Test Takeaway: The X34 OLED is slightly off the mark for HDR grayscale and EOTF tracking. Shadow detail can be restored with in-game adjustments and the grayscale error visibility will depend on what you’re looking at. Colorful games will hide the errors better and will barely look awry. If you like to play hockey sims though, the ice will be a bit green. Overall, the HDR image looks great thanks to OLED contrast, but there is a little room for improvement.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

Despite their premium pricing, OLED monitors are quickly becoming a hot category. Given the plethora of new models crossing my test bench, it’s now difficult to look in any direction without seeing one. And it’s really no surprise. They really are that good. Image quality is beyond anything an LCD, even Mini LED, can produce and game performance is next level. There is no motion blur, especially from the latest 240 Hz models, and input lag is among the lowest of any technology. And have I mentioned color gamut volume and accuracy? OLEDs do those things better too.

The Acer Predator X34 OLED is a worthy addition to the genre. It’s mostly the same as other 34-inch 21:9 WQHD 240 Hz OLEDs I’ve reviewed, but it has higher HDR brightness than all but the Asus PG34WCDM, and it has lower input lag than all but the Corsair 34WQHD240-C. For gaming, it is addictive and fun, especially when you factor in the 800R curve. That’s something few other monitors have.

I noted some issues in HDR mode where shadow detail was difficult to render. This was fixed in game by using menu tweaks to bring everything back into focus. And with no HDR calibration controls, the slight green tint I observed might be a factor if you play games with a lot of white areas. In SDR mode, the same green tint was seen out of the box, but a simple calibration took color to professional level accuracy. I also wished for a usable sRGB mode, but that was not included here.

The Acer Predator X34 OLED is a superlative gaming monitor with a very immersive quality from its 800R curved screen and stunning image. It leaves out Quantum Dots, but you save a few bucks in the process. It’s a little less expensive than the competition and it renders brighter HDR. And it’s really fast for gaming. With few flaws, it’s a great choice for those wanting a curved ultra-wide OLED monitor.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

With so much attention focused on the best gaming monitors thanks to hyper-fast screens and OLED technology, it’s always good to remember that genuinely excellent performance is available at the opposite end of the spectrum. While $1,000 will buy you a fantastic display, $162 can put a very nice 25-inch 280 Hz monitor on your desk. I’m talking about Lenovo’s Legion R25f-30, a VA panel with FHD resolution, Adaptive-Sync and wide gamut color. Let’s take a look.

Lenovo Legion R25f-30 Specs

Much of a gaming monitor’s feel comes from motion processing, which can be accomplished with high refresh rates and premium tech. The R25f-30 achieves a smooth feel with a 280 Hz overclocked panel equipped with a precise overdrive. You also get Adaptive-Sync, which works on FreeSync and G-Sync platforms, as well as consoles.  It has not been certified by Nvidia, but it ran G-Sync all day during my tests. I’ll use the phrase “surprisingly good performance” just once because the price at this writing is a paltry $162.

Though some might pause at its FHD (1920x1080) resolution, the panel size means a density of 90ppi, which is enough for a sharp image that doesn’t show its dot structure until your nose practically touches the screen. There is plenty of light output, with over 370 nits measured in my tests and around 340 for HDR content. There is no extra contrast there, but the native dynamic range for SDR and HDR is on the high end of VA territory at around 3,500:1. The picture pops, and there is plenty of color, with 90% measured coverage of the DCI-P3 gamut.

Gamers will enjoy a well-stocked OSD that has an excellent overdrive, MPRT (backlight strobe), multiple game-specific picture modes, and a frame rate counter. You also get a pair of internal speakers, but there are no USB ports. The R25f-30 also forgoes LED lighting though given the diminutive price tag, but that is no surprise.

Build quality is solid, with a fully adjustable stand and rugged construction. You also get two HDMI 2.1 ports along with a DisplayPort 1.4 and a headphone jack. The OSD is navigated by a handy joystick or by Lenovo’s Artery app that controls everything from the Windows desktop. The R25f-30 is a seriously good gaming monitor for $162. It would be easy to stop there and say “buy it,” but what fun would that be?

Assembly and Accessories

Lenovo embraces the latest recyclable packaging with a clamshell box and molded pulp insides to protect the R25f-30. It ships in three parts that assemble without tools. The panel attaches to the upright with four large knurled thumb screws. The internal power supply is fed by an included IEC cord, and you get a DisplayPort cable.

Product 360

The R25f-30 has one of the thinnest flush bezels I’ve yet seen. It’s only six millimeters wide at the top and sides and 19mm at the bottom. The anti-glare layer works well at maintaining image quality in bright rooms without introducing any grain. The word Legion appears at the lower left while a tiny LED glows at the right, white for power on and orange in standby.

The stand has Lenovo’s distinct industrial look with a girder pattern in the large base and honeycomb textures on the upright. There’s a bit of blue trim in the cable hole along with Lenovo and Legion logos silkscreened on the back. The chassis is solid and over-engineered for the R25f-30’s relatively lightweight.

Ergonomics include 5/22 degrees tilt, 30 degrees swivel and a generous 5.3-inch height adjustment. You also get a 90-degree portrait mode. Movements are very firm, as if the monitor were heavier. I expect some usage will be required to make them smoother. But the R25f-30 stays put once adjusted. You can also set the screen at eye level while making it perpendicular to the desktop.

The port panel is sparse, with just two HDMI 2.1 and one DisplayPort 1.4 plus a 3.5mm headphone jack. There are no USB ports, but you get a decent pair of internal speakers. They are limited to upper midrange frequencies but play clearly at high volumes without audible distortion.

OSD Features

The R25f-30’s OSD has clear text with a dark background for good visibility and is all business. It’s divided into five sections and appears when you click the joystick. Two additional keys take you straight to picture modes and input selection.

All OSD screens show basic signal info at the bottom. The Game Settings menu has seven picture modes, of which Standard is the default and best choice. It allows access to all calibration controls, including color temp and gamma presets. There is a four-level overdrive that can be used effectively on its maximum setting if you’re running over 240 Hz. To engage that rate, turn on the overclock option. I noted that the R25f-30 automatically reverts to 240 Hz after 30 minutes. This option cannot be defeated, which is a minor annoyance. A frame rate counter is included, but there are no aiming points. For systems running at slower frame rates, below 120fps, you can use the MPRT backlight strobe as an alternative to Adaptive-Sync. It shows a slight phasing artifact in test patterns, but in practice, it’s one of the better implementations I’ve seen.

The Screen Settings menu has basic picture controls like brightness and contrast, along with dynamic contrast, HDR (Auto or off), and a Dark Boost setting to enhance shadow detail. The R25f-30’s excellent VA panel does not need help rendering detail in any part of the image, so I recommend leaving these off. On the second screen of this menu, you’ll find Sharpness (leave it on 40) and the gamma settings. Rather than labeling them with the traditional 2.2/2.4 terms, they have correction factors. The default gamma is too light, as you’ll see in the test results, so I recommend dialing in a +0.2 correction to compensate.

If you need an sRGB mode, it is among the R25f-30’s color temp presets. All of them are fixed except the User mode which has precise RGB sliders. I used them to achieve a more precise calibration.

Lenovo Legion R25f-30 Calibration Settings

The R25f-30’s Standard mode has everything necessary for an accurate image calibration. You can use the monitor as is, but I recommend at minimum that you add a gamma correction of +0.2. The default setting is a bit light, which washes out the image. The VA panel used here is very good and has adequate contrast to support a darker gamma. With a few tweaks to the RGB sliders in the User color temp, an excellent picture can be achieved. My SDR settings are below.

HDR content signals an automatic switch if HDR is set to Auto. You can turn it off if you wish. There are no image options available, but I found reasonably close compliance with industry standards. Sadly, there is no dimming here, so contrast is the same as it is for SDR.

Gaming and Hands-on

It’s easy to be seduced by high resolution screens when playing slow-moving visual feasts such as Myst, but for high-speed combat, smooth and quick response is the most important contributor to success. Precision movement and aiming in run-and-gun maneuvers is so much easier when you have a monitor that can keep up. I’ll always choose a 240 Hz OLED, but if you don’t have $1,000 to spend, the R25f-30 is the next best thing. I don’t say that lightly. You can buy a number of monitors that run at 360 and 500 Hz. And they’ll be a little better, but not three or four times better. For gaming on a budget, this is where it’s at.

I played hours of Doom Eternal and racked up frags with ease. Classic moves like circle strafe and reverse shooting were child’s play. The R25f-30’s overdrive is super clean and removes nearly all motion blur. You won’t need an expensive PC to max the frame rate at 280. I played on a GeForce RTX 4090 equipped system and its cooling fans never turned above idle speed. Adaptive-Sync worked perfectly as well. I tried MPRT for a bit and found it effective. It didn’t remove any more blur than G-Sync with overdrive, but it would be useful for systems running below 120fps.

The picture was very colorful and deeply textured. VA panels are the next best way to create a realistic image after OLED. The R25f-30 delivers over 3,500:1 native contrast, so its black levels are deep and true, with plenty of clear shadow detail. I never found the need for the Shadow Boost option. It’s important to change the gamma setting to +0.2 to achieve the most satisfying image.

For productivity, if I had to use an FHD monitor, I would opt for this smaller size. I’ve tried many 27-inch monitors and even a few 32s and their pixel density is too low to render fine photo detail and text effectively. These things are not a problem for the R25f-30. I could not see the dot structure nor any jagged lines to distract from the workday.

I also noted the excellent sound from the internal speakers. While they won’t shake your desk with stupendous bass, they are super clean and more detailed than most. I ran the volume at 50% and could hear everything clearly without disturbing anyone in the next room. Maxing the slider gave me more than enough presence and there was no distortion.

Takeaway: The R25f-30 is an impressively capable gaming and general-use monitor. Video processing is as good as the best LCD panels I’ve reviewed, and image quality is first-rate. There’s plenty of contrast and color saturation, and pixel density is high enough to eliminate any grain or visible dot structure. For the money, it’s nearly in a class by itself.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

To compare the R25f-30’s performance, I’ve created a group of FHD monitors all close in price. Lenovo’s IPS version of this monitor, the Y25-30 is here along with AOC’s 27G15 and C27G2Z, Gigabyte’s GS27FA and GS27FC. Interestingly, I didn’t find many 25-inch screens; they were all 27, except for the other Lenovo. The latest speedy 25-inch monitors are focused on higher refresh rates, 360, 500, and 540 Hz, and therefore cost quite a bit more.

Pixel Response and Input Lag

Click here to read up on our pixel response and input lag testing procedures.

The two Lenovos dominate these tests with their 280 Hz refresh rates. On a dollar/speed basis, they are the best values available. You just need to pick your preferred screen tech, VA or IPS. In the lag test, the R25f-30 wins the day with a very quick 18ms total score. That’s on par with many costlier monitors that run at higher refresh rates. The R25f-30 is further enhanced by its precise overdrive that removes nearly all traces of motion blur.

Test Takeaway: The R25f-30 and Y25-30 are two of the quickest monitors you can buy for less than $200. They both have excellent overdrive, high motion resolution and extremely low input lag. For competition, you’ll need to spend significantly more money to beat them. Realistically, the only way you’ll achieve a better gaming feel is with an OLED.

Viewing Angles

Viewing angles are not VA’s strong suit, but the R25f-30 acquits itself well, especially considering its low price. The side view stays bright with no visible light loss and shifts slightly to red and green. Gamma becomes lighter, which slightly washes out the picture. The top view is almost entirely void of detail, with a 40% light reduction and a red tint.

Screen Uniformity

To learn how we measure screen uniformity, click here.

A slight hotspot in my R25f-30’s upper left corner spoiled the uniformity test result. I could see a bit of glow in a full black pattern with the room lights turned off. It wasn’t a factor in content unless the entire picture was near-black. This is a sample-specific issue.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

To read about our monitor tests in-depth, please check out Display Testing Explained: How We Test PC Monitors. We cover brightness and contrast testing on page two.

Uncalibrated – Maximum Backlight Level

The R25f-30 is one of the brighter VA monitors I’ve tested, with a max output of 373 nits in SDR mode. That’s more than enough light for any indoor environment. Few users will need to set brightness above 200 nits so there is plenty of headroom for HDR content and its specular highlights.

Though it takes third place in the contrast contest behind the 27G15 and GS27FC, the R25f-30 is still above the VA average with a respectable 3,737.4:1 ratio thanks to its low black level.

After Calibration to 200 nits

Calibration doesn’t change much except for the output level equaling 200 nits. The Gigabyte and AOC screens have a slight edge but, in a side-by-side comparison, you’ll be hard-pressed to see a difference. In the ANSI test, the R25f-30 is within a whisker of the numbers posted by the top two screens. That’s definitely a photo finish. With over 3,000:1 intra-image contrast, the R25f-30 is one of the better VA panels I’ve tested.

Test Takeaway: The R25f-30 has plenty of light output and overhead for HDR content. Its contrast is above the VA panel average and image quality is well above what you’d expect for the price. None of these monitors are expensive, but the R25f-30 delivers champagne quality for little more than beer money.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

The R25f-30 delivers decent color accuracy in its Standard mode. You can use it without changing the color temperature, but it benefits from a gamma tweak, which I’ll talk about now.

Grayscale and Gamma Tracking

Our grayscale and gamma tests use Calman calibration software from Portrait Displays. We describe our grayscale and gamma tests in detail here.

The R25f-30’s default grayscale runs a bit warm with a slight brown tint. It would be great for watching old movies if you like the sepia tone. The error level is just at the point of visibility from 60% to 100% brightness. The greater issue is gamma which is clearly too light. This washes out the image and robs it of liveliness.

Changing the gamma offset to +0.2 makes a world of difference. You could do that and see an improvement but going further with RGB adjustments takes care of any color tint you might see. Calibrated grayscale is in the reference zone with excellent gamma tracking.

If you need sRGB color, that option is available in the color temp menu, but it is not adjustable. Its grayscale tracking is warm in tone with an obviously light gamma. As you’ll see later, its color is under-saturated as well.

Comparisons

The R25f-30’s 3.42dE grayscale error is just over the visible threshold. It isn’t too obvious, but correcting it will make a positive impact. The final 0.83dE score is one of the better results for monitors costing less than $200. Gamma tracking shows a similar result and ranking with a 0.14 range of values and an average of 2.26, 2.73% off the 2.2 standard. Dark gamma is well suited to VA panels thanks to their high native contrast.

Color Gamut Accuracy

Our color gamut and volume testing use Portrait Displays’ Calman software. For details on our color gamut testing and volume calculations, click here.

The R25f-30 has excellent out-of-box gamut accuracy and significant coverage of DCI-P3. This is another impressive point when considering its low price. You won’t find much better color than this without spending more than $200. The default chart shows near-perfect red compliance with some hue errors in cyan and magenta. Blue is also a little under-saturated.

Calibration fixes these issues and adds a bit of overall saturation thanks to the change in gamma. Only green comes up a tad short, but that’s a typical result for wide gamut displays. It certainly doesn’t get better than this for under $200.

The sRGB result is disappointing. Red and blue are under-saturated, and magenta is off hue. There are no calibration options here, so you get what you see. If you really need an accurate sRGB mode, there are better choices than the R25f-30.

Comparisons

With a final color score of 1.53dE, the R25f-30 ranks among the best. This is pro-level performance which is only approached by the C27G2Z in second place. The R25f-30 also has very high coverage of the DCI-P3 gamut at over 90%. It’s a couple of points ahead of the next screen and only a hair behind the 27G15. Further down the chart, you can see more typical gamut volume performance for this price point. The R25f-30 truly excels.

Test Takeaway: The R25f-30 delivers far better color accuracy and volume than one would expect for $162. It isn’t begging for calibration, but a few tweaks will have a positive impact on picture quality. The default gamma is a bit light, but that’s an easy fix. For the money, it doesn’t get much better than this.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

Our HDR benchmarking uses Portrait Displays’ Calman software. To learn about our HDR testing, see our breakdown of how we test PC monitors.

The R25f-30 supports HDR10 signals with an automatic switch to HDR mode. Set the HDR option to Auto to ensure this. There are no picture adjustments available.

HDR Brightness and Contrast

The R25f-30’s HDR peak output is slightly lower than I measured for SDR. It’s still high enough to bring out specular highlights like metallic reflections or lens flare. Unfortunately, there’s no zone or field dimming to increase dynamics. The black levels and contrast are therefore the same as what I saw for SDR. In fairness, none of these monitors have dimming options so the Lenovo is on equal footing. And it benefits from its VA panel and high native contrast. You can see that when comparing it to the IPS panels in fifth and sixth place.

Grayscale, EOTF and Color

Viewing a grayscale step pattern in the R25f-30’s HDR mode shows a red tint which is supported by the test result. Red errors are visible above 40% brightness. It isn’t a huge negative because it’s less visible than blue or green issues would be. But it would be nice to have some color temp options here. The EOTF tracks under the reference line which means it’s a little too dark. The linearity is there so detail remains visible. And the lowered black levels add a bit more punch to the image. At this price point, I can’t complain too loudly.

In the HDR color test, there is general over-saturation in the primary colors which is what I see from most wide gamut displays. This also adds impact to where the HDR image is a little livelier than the SDR one. The highest green points are a little under-saturated, but that is barely a factor in actual content. The R25f-30 makes a good attempt at Rec.2020 with over-saturation of the inner points. It runs out of color at 85% red, 65% green and 90% blue. This is also typical of the wide gamut displays I’ve tested.

Test Takeaway: The R25f-30 does a good job with HDR content in the context of displays that have no dimming feature. Contrast is the same for SDR and HDR but since the R25f-30 is a VA monitor, it outperforms its IPS competitors. Color is reasonably accurate so your HDR games will get a small boost in quality. This is as good as it gets for value-priced monitors.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

There are many good gaming monitors at both extremes of the price spectrum. If you have $1,000 to spend, there are fast OLEDs that will satisfy. But if you have less than $200, you can get much of that experience. For good gaming, you need speed and smoothness first, followed by high contrast and saturated color. The Lenovo R25f-30 has all these for $162 at this writing.

At 25 inches, the FHD resolution is not a hindrance. Pixel density is 90ppi, so you’ll get smooth font edges and good detail in your favorite games. You can edit photos too, as long as you don’t need sRGB color. The R25f-30 falls a bit short there in both volume and accuracy. But it covers more of DCI-P3 than most budget screens and delivers good accuracy out of the box and great accuracy with a simple calibration. And with higher-than-average VA contrast, it has better picture depth than most sub-$200 displays.

Gaming performance can only be called superb. The overdrive is what I’d expect from a more expensive screen with precise operation at 280fps and near-complete elimination of motion blur. The MPRT backlight strobe is a viable option for users running below 120fps. And Adaptive-Sync worked in all instances without a hiccup. The R25f-30 is a precise and reliable gaming tool for fast-paced competition. And I noted better than average sound quality from the internal speakers. Most monitors treat these as an afterthought, but Lenovo has paid more attention. They are clear, loud and well-detailed.

The Lenovo Legion R25f-30 delivers an astounding gaming experience for just $162 at this writing. You’ll need to spend three or four times as much to get a little more performance. It’s one of the best budget screens I’ve reviewed of late, which is excellent news for bargain shoppers.

MORE: Best Gaming Monitors

MORE: How We Test PC Monitors

MORE: How to Buy a PC Monitor

MORE: How to Choose the Best HDR Monitor

The Crucial T500 4TB is here, for those who find that 2TB of storage just isn’t enough these days. Using Phison's E25 controller, the T500 has a power-efficient architecture with DRAM, now available as a 4TB pseudo-budget SSD. Some of our favorite PCIe 4.0 drives simply haven't bothered with higher capacities, like the SK hynix Platinum P41, so attention has turned to the T500. Crucial promised a 4TB model a year ago, and it has now delivered: the same great taste but in an extra large tin.

For the most part, nothing has changed but the capacity. The T500 was heralded as a top notch PCIe 4.0 SSD that could match or beat any other drive in its class, all while being more power-efficient. Half the channels with the same amount of fun, and unlike some of its competitors, it kept the DRAM intact. It’s a great successor to drives like the SK hynix Gold P31 for laptops, and it also works great for desktops and consoles. What’s not to love?

Perhaps the most obvious potential misstep is the price. In some respects this is supposed to be a “budget” SSD, even though its performance is in line with heavy hitters. Often, it manages to look like a WD Black SN850X or Samsung 990 Pro substitute, but really it just brings DRAM to a potentially less expensive drive, taking on the likes of the Teamgroup MP44. This becomes more obvious when looking at its sustained write performance irregularities, although for the most part you’re not going to notice that in daily use. Still, its price is higher than it should be at 4TB, which keeps it from being the best option in many cases, even if it delivers exemplary all-around performance.

Crucial T500 Specifications

The 4TB Crucial T500 is currently priced at $309 for the non-heatsinked version and $301 for the heatsink model. Those prices can and will change — a week or two back, the non-heatsink model was going for $289 and the heatsink variant cost $319. Whether you want the heatsink or not, check the current pricing, as you can easily remove the heatsink if you really don't want it. If the bare drive ends up costing less (as it logically should), a variety of M.2 heatsinks will get the job done, or you can use your motherboard's M.2 heatsink if it has one. Do note that the 4TB model isn't a single-sided design, however.

The Crucial T500 4TB can hit up to 7,000/6,900 MB/s for sequential reads and writes, and up to 1,050K/1,350K random read and write 4K IOPS. These numbers are slightly lower than the 2TB SKU, and the 1TB SKU as well for that matter. The reason is the hardware. The controller in use has four chip enable (CE) signals for each of the four channels, giving sixteen CE total. As the drive is using 1Tb dies there are 32 dies in total present, more than the CE count, which add some overhead. We’ll point out physical layout changes below that relate to this configuration. If you’re concerned about performance loss, we would state that these dips are not significant in everyday use and that this is a capable 4TB SSD.

The 4TB T500 has a warranty for five years and up to 2,400TB of writes, which is right in line with the smaller SKUs. The drive also supports TCG Opal 2.01 for hardware encryption.

Crucial T500 Software and Accessories

The 4TB T500 has the same software available as discussed in the original Crucial T500 2TB review — Crucial’s Storage Executive SSD toolbox and Acronis True Image. The toolbox has typical features including health management, secure erase, firmware updates, and encryption support. Acronis True Image is for imaging, cloning, and backing up your data, which is useful especially when upgrading your primary drive.

Crucial T500: A Closer Look

One of the biggest questions we had was whether the 4TB T500 would be single-sided. The answer, unfortunately, is no. It’s theoretically possible to stack 16 dies per package (16DP) and 1Tb flash is in abundance, but there are reasons to prefer four 8-die packages (8DP/ODP). Due to the T500’s layout, which includes a diagonal and sizable controller as well as a package of DRAM, the drive must be double-sided at 4TB when using 8DP.

Plenty of other companies have put 4TB on a single-sided drive, though often that's with DRAM-less drives. But the Samsung 990 Pro 4TB as an example puts four 1TB NAND packages on one side of the stick, and that can be better for cooling in a laptop or console. The 45-degree rotation on the Phison E25 controller would already make it difficult to do four NAND packages on a side.

But the flash density and performance also requires additional ICs on the PCB that we'll discuss below, this is unfortunately the easier way for Crucial to go, and results in a double-sided form factor. This might preclude the drive’s use in some systems at this capacity, although most laptops today can handle double-sided drives.

The hardware is effectively the same as before with the Phison E25 controller, LPDDR4, and Micron’s 232-Layer TLC flash. One thing that stands out is the presence of additional ICs listed as DDR4000 by Texas Instruments. This is not the first time we’ve seen these. They were also present on the Corsair EX100U and Sabrent Rocket nano V2 portable drives.

On this drive there are a total of four of these ICs, one for each of the E25’s four channels. These could be used for many things, but on SSDs they are commonly used to keep signaling within acceptable parameters for the 2,400 MT/s flash at hand. This becomes more of an issue as more flash is added, and for 4TB — which requires 32 dies, or 8 dies per channel — this is a sure way of maintaining performance with the extra loading. The previous drives we tested had these at 2TB with slower flash, but those drives were only using two-channel controllers. The use of these ICs on the 4TB T500 further encourages a double-sided design due to space constraints.

MORE: Best SSDs

MORE: How We Test HDDs And SSDs

MORE: All SSD Content

Comparison Products

We’re starting out by looking to the future with the PCIe 5.0 Crucial T700 and at the past with the early adopter Sabrent Rocket Q4. Moving forward, 4TB drives will become more commonplace where once you looked to QLC flash to make things work. The T500 will inevitably be compared with top-tier 4.0 drives like the Samsung 990 Pro — which updated its hardware with the 4TB launch, and the drive remained single-sided — but is more fairly compared to Phison E18-based legends. These drives used the first high-end 4.0 hardware available and remain good picks, particularly for sustained workloads, albeit with higher power use and worse efficiency. These include the Sabrent Rocket 4 Plus-G, the Seagate FireCuda 530, the Teamgroup Carda A440 Pro, and the Netac NV7000.

The T500 also has to face off against other budget 4TB drives, though these are DRAM-less designs. The Addlink A93 and Lexar NM790 remain popular choices with often lower price tags. Or, you can go with QLC flash to save some money, as with the Patriot VP4300 Lite or more entry-level WD Blue SN5000. QLC flash still has a poor reputation, but for extra inexpensive storage it remains a reasonable compromise. 

Trace Testing — 3DMark Storage Benchmark

 Built for gamers, 3DMark’s Storage Benchmark focuses on real-world gaming performance. Each round in this benchmark stresses storage based on gaming activities including loading games, saving progress, installing game files, and recording gameplay video streams. Future gaming benchmarks will be DirectStorage-inclusive and we also include notes about which drives may be future-proofed.

The T500 is the fastest PCIe 4.0 SSD on the market when it comes to 3DMark. This does not translate to significantly faster loading times or higher FPS, but it does mean it makes a great game drive. Its controller also has Phison’s DirectStorage-optimized I/O+ firmware, so it’s future-proofed for upcoming games. Drives built on Phison’s newer E31T controller will have higher bandwidth and power efficiency, but the T500 still stands out by having DRAM. This could make for more consistent performance with DirectStorage titles, but that remains to be seen.

Trace Testing — PCMark 10 Storage Benchmark

PCMark 10 is a trace-based benchmark that uses a wide-ranging set of real-world traces from popular applications and everyday tasks to measure the performance of storage devices. The results are particularly useful when analyzing drives for their use as primary/boot storage devices and in work environments. 

The T500 is also the fastest 4.0 drive in PCMark 10. It’s an excellent all-around drive that can work well as the sole drive in your PC — or any system for that matter. For laptops, PCIe 5.0 SSDs remain out of reach, although a few built on the E31T are coming very soon. The T500 remains competitive, though, thanks again to its DRAM, the presence of which some people prefer. This makes the T500 the natural SK hynix Gold P31 successor and more or less irreplaceable for the time being if DRAM’s your thing. Application performance should be excellent across-the-board. 

Console Testing — PlayStation 5 Transfers

The PlayStation 5 is capable of taking one additional PCIe 4.0 or faster SSD for extra game storage. While any 4.0 drive will technically work, Sony recommends drives that can deliver at least 5,500 MB/s of sequential read bandwidth for optimal performance. In our testing, PCIe 5.0 SSDs don’t bring much to the table and generally shouldn’t be used in the PS5, especially as they may require additional cooling. Check our Best PS5 SSDs article for more information.

Our testing utilizes the PS5’s internal storage test and manual read/write tests with over 192GB of data both from and to the internal storage. Throttling is prevented where possible to see how each drive operates under ideal conditions. While game load times should not deviate much from drive to drive, our results can indicate which drives may be more responsive in long-term use. 

Our performance results for the 4TB T500 on the PS5 are a little lower than we expected. In our original T500 review we did notice some discrepancies in our PS5 testing, but Crucial was aware of this and later put out a firmware update. The 4TB model comes with new firmware pre-applied, but the PS5’s firmware has also since been updated in a way that does reduce our scores a little bit on all drives. Even with this considered, the 4TB T500 underperforms to a small degree, but not enough that you shouldn’t consider it very usable for the console. Any relatively recent PCIe 4.0 SSD should generally work fine in the PS5, with price per GB often being the determining factor.

Transfer Rates — DiskBench

We use the DiskBench storage benchmarking tool to test file transfer performance with a custom, 50GB dataset. We write 31,227 files of various types, such as pictures, PDFs, and videos to the test drive, then make a copy of that data to a new folder, and follow up with a reading test of a newly-written 6.5GB zip file. This is a real world type workload that fits into the cache of most drives. 

Our DiskBench data fits in just about any pSLC cache these days, so the results here are an attempt to show real world transfer performance before the cache runs out. For larger, sustained transfers, see our write saturation testing below. We mention this because the T500 performs excellently in DiskBench but still has some issues when its cache gets depleted. We do think that, in the real world and especially with the size of the cache for the 4TB T500, this drive should perform excellently in almost all cases.

Synthetic Testing — ATTO / CrystalDiskMark

ATTO and CrystalDiskMark (CDM) are free and easy-to-use storage benchmarking tools that SSD vendors commonly use to assign performance specifications to their products. Both of these tools give us insight into how each device handles different file sizes and at different queue depths for both sequential and random workloads. . 

The T500’s ATTO performance is good, with only two dips worthy of notice. The first is with sequential reads at the 2MiB block size. This is possibly because of a boundary that exists with flash interleaving, where essentially you can have four six-plane dies for each of four channels for peak performance. Given the typical plane size of 16KiB, this may create a boundary between 1MiB and 2MiB. In the real world, this is not a significant issue.

There are also dips at 2KiB for both reads and writes, although not nearly as bad as with the T700. While consumer flash today has 16KiB physical pages, it’s not uncommon to have these broken down into 4KiB logical pages. This matches typical cluster and physical sector sizes, although a logical sector size of 512B is normal. This would leave 2KiB I/O as a subpage, which has interesting ramifications for the SSD controller and its DRAM mapping memory. This performance dip is likely deceptive as Phison’s work with enterprise storage, where 2KiB I/O may need optimized handling, informs its consumer controller design. As the T500 has plenty of DRAM and ideally you have some queue depth for that type of I/O, this result may be an anomaly.

CDM gives us a fuller picture by revealing random 4KB QD1 performance — the T500 is very good here — as well as sequential throughput. The T500 is as fast as any other high-end PCIe 4.0 SSD, so no surprises there. Given the relatively minute differences, you’re not missing anything by going with the T500. You can safely buy it expecting more or less the best performance possible for a drive of its class. 

Sustained Write Performance and Cache Recovery

Official write specifications are only part of the performance picture. Most SSDs implement a write cache, which is a fast area of pseudo-SLC (single-bit) programmed flash that absorbs incoming data. Sustained write speeds can suffer tremendously once the workload spills outside of the cache and into the "native" TLC (three-bit) or QLC (four-bit) flash. Performance can suffer even more if the drive is forced to fold, which is the process of migrating data out of the cache in order to free up space for further incoming data.

We use Iometer to hammer the SSD with sequential writes for over 15 minutes (we did two hours in this case) to measure both the size of the write cache and performance after the cache is saturated. We also monitor cache recovery via multiple idle rounds. This process shows the performance of the drive in various states as well as the steady state write performance. 

In our original review of the T500 we noticed some inconsistencies in write performance, particularly when the pSLC cache was exhausted. Crucial later put out a firmware update intended to improve in this area, but in our testing things were still relatively irregular. The 4TB, with its relatively large cache, still seems to suffer in this area, although technically speaking the base write performance of the TLC flash is excellent. This is revealed in the steady state write speeds which are supposed to yield the native flash performance, more or less. 3.5 GB/s for a four-channel controller like the T500’s would be exceptionally fast.

The 4TB T500 begins by writing at around 6.5 GB/s in the single-bit pSLC mode for up to almost five minutes. This is a huge cache and is one of the reasons post-cache performance looks so slow. The drive is eventually forced to empty the cache to free up the full amount of space and this naturally bottlenecks things. In this slower folding mode, where the drive is forced to copy data from pSLC to the native TLC flash while also juggling incoming writes and migrating blocks from pSLC to TLC mode, the drive writes at under 600 MB/s. This is rather sluggish, but far from the worst we’ve seen. QLC-based drives even at 4TB can only achieve half this speed or lower.

All of this makes for some deceptive results. We know the Phison E25 controller is designed for up to about sixteen dies of interleaving, given that it’s a four-channel and not eight-channel design. We also know how fast this flash is, so realistically it wouldn’t plod away at a 3.5 GB/s average, either. A more reasonable target would be what we see on the A93 and NM790, which uses YMTC’s comparable 232-Layer TLC flash. That result would still be very good. We’re instead seeing an average of the pSLC and folding modes, which added together and divided by two is 3.5 GB/s or so.

So, what does this mean? Well, the T500 is simply not very consistent when dealing with larger writes. If the drive is fuller the cache will be smaller so the size of the transfer required to make this happen will change. For the most part, you should never see poor performance from this drive as the cache does a good job for 99% of workloads. You shouldn’t be buying this drive for anything super intense, like massive sustained writes. That said, these results can make people nervous, in part because many drives in the past have had pSLC issues in the long-run. Getting “stuck” in TLC mode isn’t as bad as it looks with our results, though, but we can understand why you might want something that performs more consistently here. 

Power Consumption and Temperature

We use the Quarch HD Programmable Power Module to gain a deeper understanding of power characteristics. Idle power consumption is an important aspect to consider, especially if you're looking for a laptop upgrade as even the best ultrabooks can have mediocre stock storage. Desktops may be more performance-oriented with less support for power-saving features, so we show the worst-case.

Some SSDs can consume watts of power at idle while better-suited ones sip just milliwatts. Average workload power consumption and max consumption are two other aspects of power consumption but performance-per-watt, or efficiency, is more important. A drive might consume more power during any given workload, but accomplishing a task faster allows the drive to drop into an idle state more quickly, ultimately saving energy.

For temperature recording we currently poll the drive’s primary composite sensor during testing with a ~22°C ambient. Our testing is rigorous enough to heat the drive to a realistic ceiling temperature.

The 4TB T500 is slightly worse in all metrics when compared to our original 2TB sample: higher idle, peak, and average power consumption, with somewhat lower overall efficiency. This isn’t an exact comparison as there have been firmware updates since then, so we did not include the original results. While packing in more flash can reduce power efficiency at some point, modern drives are very good at switching dies to use as little power as is possible. The 4TB does have the DDR switch/multiplexers — NAND flash runs on some similar principles as DRAM memory with, say, NV-DDR3 — which could eat a small amount of power. Regardless, the drive is still fairly efficient, particularly as it has DRAM, and power consumption is not a huge concern.

In our testing the drive hit a maximum temperature of 57°C. This is well below the throttling threshold, so any well-cooled system should have no problems running this drive. It will also be fine in a laptop or PS5. You can definitely add a heatsink, though — or buy the model that comes with one — for peace of mind. This is not a drive that should get too hot as you'll want to avoid massive sustained workloads anyway.

Test Bench and Testing Notes

We use an Alder Lake platform with most background applications such as indexing, Windows updates, and anti-virus disabled in the OS to reduce run-to-run variability. Each SSD is prefilled to 50% capacity and tested as a secondary device. Unless noted, we use active cooling for all SSDs.

Crucial T500 Bottom Line 

If you’ve been looking for a 4TB T500, here it is. Crucial always promised to offer this capacity — a capacity that’s often clamored for, since it feels like many drives top out at 2TB these days — and it has delivered. It’s a power-efficient drive with DRAM, which is part of what drives the desire to get it at 4TB. Great performance for a laptop drive? Sounds like an amazing product. Although, this drive is double-sided and has some other caveats that keep it from being the perfect dream come true.

All-around performance remains excellent, and in fact it’s the fastest PCIe 4.0 SSD you can get when all things are considered. Throughput and latency are both top notch. This comes with higher power consumption than equivalent DRAM-less drives, but it has DRAM. That’s a fair trade-off, but what isn’t is the drive’s sustained performance. Despite firmware updates, the T500 is still inconsistent when it comes to longer sustained-write workloads. Its large cache should hide this well for all but the most hardcore users, but it remains a sticking point that keeps this drive from being flawless. If you need DRAM and sustained performance, you’re better off getting a 4TB drive with an eight-channel controller: the WD Black SN850X, the Samsung 990 Pro, or an E18-based drive like the Sabrent Rocket 4 Plus-G.

Crucial sells the T500 both with and without a heatsink, which is a great option to have for certain builds. This is something else that makes it a great all-around drive, but ultimately it falters with pricing at 4TB. At the time of review it’s still being sold at $50+ over decent DRAM-less drives like the Lexar NM790 or Teamgroup MP44 which makes it a hard sell unless you really want DRAM. That begs the question: is DRAM worthwhile?

This is an awkward question to answer because areas where DRAM is especially useful — with a fuller drive, running heavier workloads, that sort of thing — the T500 can be less consistent due to its cache handling. For everyday use, on the other hand, you probably wouldn’t be able to blind guess which drive has DRAM. It’s more of a nice-to-have.

One thing that challenges the T500’s relevance is the fact that its primary competitors, such as drives that use the DRAM-less Phison E27T — which is a possible sidegrade for the MP44 and others, and the default for the Corsair MP600 Elite and others — are excellent choices for laptops even over something like the NM790. These drives run cool with high performance and none of the sustained write shenanigans. The problem is, none of these are available at 4TB. What is available are 4TB drives with QLC instead of TLC flash, if that’s a compromise you’re willing to make. And it's a reasonable option if you just need extra storage in a secondary drive at the lowest possible price.

If you have your heart set on that much storage with TLC flash, though, the T500 is a top runner for now, even with its high relative cost. But it does have some real direct competition from existing drives, like the Samsung 990 Pro 4TB and WD Black SN850X 4TB, both of which also cost $309 at the time of writing. Depending on the workload, the T500 can come out ahead of or behind those drives (though we never tested the SN850X 4TB, we did just recently check out the SN850X 8TB).

Overall, the T500 is a good achievement. It’s basically a modern replacement for the SK hynix Gold P31 — or as a higher capacity option than the SK hynix Platinum P41. It’s faster and more efficient than the Gold P31 and is now available at up to 4TB. If you’re looking to upgrade your laptop from an older or OEM drive, or even a Gold P31, the T500 looks pretty good. With pricing the way it is, though, it’s probably not the best PS5 option. For desktop use, it’s somewhat of a push depending on what you’re trying to do. If money is no object and you just want an excellent all-around drive, it’s arguably the best drive on the market, and we can highly recommend it. 

MORE: Best SSDs

MORE: How We Test HDDs And SSDs

MORE: All SSD Content

Nvidia is clamping down on playtime for players of its GeForce Now cloud gaming service. The trillion-dollar company announced it will introduce a "100-hour monthly playtime allowance" at the start of next year. Hardcore gamers who might hit the 100-hour limit can extend their playtime with a small fee.

This change is purportedly designed to help keep GeForce Now's "exceptional" quality and speed consistent and provide short queue times specifically for Performance and Ultimate players. Nvidia says the 100-hour limit is generous, with only 6% of its entire GeForce Now user base actually being impacted by the change. 100 hours of playtime gives players three hours per day of game time if they play every day for an entire month.

For anyone who manages to hit the 100-hour monthly limit, Nvidia is charging its customers $2.99 for an additional 15 hours of playtime for users on its Performance tier and $5.99 for the same playtime extension for Ultimate tier players for anyone who wants to bypass Nvidia's 100-hour limit. Nvidia will provide the ability to monitor your playtime in GeForce Now so users can monitor how close they are (or aren't) to the 100-hour limit.

Nvidia has also announced some upgrades to its current GeForce Now membership plans. GeForce Now's previous "Priority" plan has been replaced with the "Performance" mid-tier plan, which sees an upgrade to 1440p. This enables gamers to play games at up to 1440p native resolutions; previously, gamers were limited to 1080p on the Priority plan. In-game graphics settings can also be saved during streaming sessions. The Ultimate and free plans see no change.

As a thank you to its concurrent player base, Nvidia is offering active paid members, as of December 31 of this year, the ability to continue with unlimited playtime for all of 2025. This effectively delays the 100-hour monthly limit by a full year for users with an active plan as of December 31, 2024, or earlier.

Nvidia is also offering a temporary 25% discount on its Day Passes. The Performance Day Pass will drop to $2.99, and the Ultimate Day Pass will drop to $5.99. Day passes offer 24 hours of access to either the Ultimate or Performance plans before you have to pay again, switch to a monthly subscription, or revert back to the free plan.

AMD’s $480 Ryzen 7 9800X3D comes armed with eight cores and 16 threads paired with a new version of the company’s game-boosting 3D V-Cache tech that delivers impressive performance, taking the throne as the fastest gaming CPU on the market in convincing fashion. AMD claims the chip is 20% better than Intel’s latest flagship and 8% faster than the prior-gen Ryzen 7 7800X3D. However, our benchmarks show it beats Intel’s current-gen flagship Core 9 285K by an almost unrealistic 35% on average in our test suite. It even beats Intel’s fastest competing gaming chip, the Core i9-14900K, by an equally incredible 30% on average in our test suite. Naturally, the gains will vary by title due to the vagaries of the 3D V-Cache tech. Still, we recorded incredible performance that firmly cements the Ryzen 9 9800X3D as the best CPU for gaming.

AMD accomplished this feat by leveraging its tried-and-true 3D V-Cache technology, which uses a beefy, vertically-stacked 64MB L3 cache chiplet to boost gaming performance. To improve performance further, AMD moved the cache chiplet from the top of the processor to the bottom. This yields substantially higher thermal headroom, enabling higher clock rates while the additional cache unlocks the Zen 5 architecture’s peak gaming performance. As we're accustomed to with gaming-optimized X3D processors, not all games will benefit from the tech, and a few are even slightly slower (1 to 2%) than the previous-gen model. Still, the chip also vastly improves 1% low frame rates even in those titles. The higher 1% lows will also help when you’re GPU-limited, like when gaming with higher resolutions.

AMD has also now enabled full overclocking for the 9800X3D, which we’ve put to the test, allowing you to scrape out more performance in both gaming and productivity work. Speaking of which, the new design also helps to remove some of the heavy penalties in productivity work that you’ll often pay for using a gaming-optimized X3D processor, allowing the 9800X3D to match, and at times even exceed, a comparable eight-core Ryzen 9000 CPU in several productivity workloads.

The $480 Ryzen 9 9800X3D is AMD’s lone chip launch today, and the chip will be available on shelves tomorrow. The 9800X3D saves the day for gaming enthusiasts faced with not just one, but two disappointing chip launches this year: AMD’s Zen 5-powered Ryzen 9000 series debuted with lackluster gaming performance gains when it arrived earlier this year, but as you’ll see in our fresh testing we did for this review, recent firmwares and a Windows tweak have wrung out more performance. Intel also released its highly anticipated Arrow Lake Core Ultra series just two weeks ago, but those chips failed to match the gaming performance of Intel’s prior-gen chips. Unfortunately, there’s no evident fix in sight, which set the stage for a disappointing coming year for gaming chips.

The Ryzen 7 9800X3D changes that completely, bringing a top-tier gaming chip to market at a reasonable price point given its capabilities. The chip drops into existing AM5 motherboards and won’t require the pricey complementary components, like the pricey high-speed RAM kits, beefy coolers, and motherboards needed to extract the utmost performance from other top-tier chips. Meanwhile, Intel seems wholly unable to counter AMD’s 3D V-Cache tech even after two and a half long years and now three chip generations, after it debuted with the now-legendary Ryzen 7 5800X3D. Let’s look at the tech, then move on to the surprisingly good gaming benchmarks.

AMD Ryzen 7 9800X3D Pricing and Specifications

The Ryzen 7 9800X3D features the Zen 5 microarchitecture, which you can read more about here. The processor is compatible with AM5 motherboards, and AMD's partners finally have X870E models available on the market. 

The eight-core, 16-thread Ryzen 7 9800X3D debuts at $480, a $30 premium over the launch pricing of the Ryzen 7 7800X3D. The 9800X3D has a 120W TDP and features 500 MHz higher base clocks and 200 MHz higher boosts than the previous-gen Zen 4-powered Ryzen 7 7800X3D.

The 9800X3D's 5.2 GHz boost is impressive, but that lags the direct Zen 5 comparable, the eight-core Ryzen 7 9700X, by 300 MHz. However, AMD has compensated with a 900 MHz higher base clock, allowing the 9800X3D to outpace the 9700X in some threaded workloads, as you'll see in our benchmarks on the following pages.

As with the prior-gen models, the Ryzen 7 9800X3D doesn't come with a cooler. AMD recommends at least a 240-280mm liquid (or equivalent) cooler. The processor has the same 95C maximum temperature (TjMax), but the new L3 cache chiplet design allows the chip to operate at higher clock rates for longer periods of time (enhanced boost residency), which equates to stronger performance gains within the same TDP envelope.

The Ryzen 7 9800X3D is fully overclockable — you can tune the CPU cores, fabrics, and memory to your liking. However, while multiplier-based overclocking is available, as we've seen with other Zen 5 processors, most users with conventional cooling will be best served using the auto-overclocking Precision Boost Overdrive (PBO) feature. We have extensive testing with this feature enabled on the following pages.

AMD also bumped memory support up to DDR5-5600 from the DDR5-5200 found with the previous-gen 7800X3D.

AMD Ryzen 7 9800X3D architecture

In the illustrations above, you can see AMD's original approach with its 3D V-Cache tech. With previous designs that leveraged the Zen 3 and Zen 4 architectures, AMD stacked an additional L3 SRAM chiplet directly in the center of the compute die (CCD) chiplet to isolate it from the heat-generating cores. We’ve covered the details of the first generation of this technology here.

AMD used hybrid bonding technology to join the cache chiplet to the underlying compute die. This boosted cache capacity to 96MB to accelerate performance in latency-sensitive applications, like gaming.

AMD placed the chiplet and several pieces of structural silicon on top of the compute die, but as we demonstrated with thermal throttling tests of the Ryzen 9 7950X3D processor, the cache chiplet and structural silicon shims served as a thermal blanket that hampered effective thermal transfer, essentially trapping waste heat. As a result, AMD limited the effective voltage and frequency range of the 3D V-Cache-equipped die to rein in heat generation. Simply put, the chip ran hot and thus had to run slower. 

AMD has now placed the cache chiplet underneath the die to alleviate the thermal challenges, placing the compute die closer to the thermal interface material, IHS, and eventually the CPU cooler. This improves the amount of compute performance AMD can extract from the 120W TDP envelope, and it also improves boost residency/duration. However, it comes with a new set of challenges — this design requires AMD to route the power and signal TSVs for the entire chip through the underlying L3 cache chiplet.

The new L3 cache chiplet is based on the same 7nm SRAM-optimized process node that AMD used in the previous two generations of 3D V-Cache technology in a technique that AMD now refers to as its 'Second-Gen 3D V-Cache' technology (this is an odd branding choice — this is actually the third refinement of the technology).

As before, AMD thins both the L3 cache chiplet and the 4nm (N4P) compute die to adhere to the standard Z-Height requirements for the processor package. AMD uses the same hybrid copper-to-copper bonding with a 9-micron bump pitch to join the two die. 

The chiplet now spans the entire bottom of the compute die, thus eliminating the need for the structural silicon shims. AMD isn't sharing the size of the L3 cache chiplet or the transistor density, but the compute die measures 70.6 mm^2, so it's safe to bet that's also the size of the cache chiplet. That's much larger than the 36 mm^2 L3 cache chiplet used for the Zen 4 X3D models — this is despite the use of the same process node and the same 64MB of L3 capacity. AMD likely has a lot of 'empty' silicon on the cache chiplet that's simply there for structural support, without the need to add separate shims.

Naturally, AMD has to feed more power and signal TSVs through the L3 cache chiplet, which now serves as an interposer of sorts, up to the compute die. AMD says that power TSVs are distributed throughout the SRAM die to feed the compute die that resides above. Extra power TSVs are distributed in areas that are unused for other functionality (presumably as more cache storage), which should at least partially account for the larger size. 

AMD says accesses to the L3 cache chiplet incur the same four-clock penalty as with the prior generation. We measured a sub-ns increase in L3 latency compared to the prior-gen 3D V-Cache-equipped Ryzen 7 7800X3D, which falls within the expected variance. AMD says the bandwidth between the compute die and L3 chiplet is similar to the previous-gen at 2.5 TB/s, but bandwidth varies with clock speeds.

Accesses to both main memory and I/O now have to travel through the L3 cache chiplet. As you can see in the measurements above, we see a marked increase in DDR5 latency over the standard Ryzen 7 9700X and the Ryzen 7 7800X3D. AMD says there is no increase in I/O latency due to traveling through the cache chiplet, though. We're following up for more details.

Naturally, any increase in memory latency is undesirable. However, we expect this will have minimal impact on overall performance due to the larger slab of L3 cache that minimizes accesses to main memory in some scenarios. Let's see what performance looks like in gaming on the next page. 

• MORE: Best CPUs for Gaming
• MORE: CPU Benchmark Hierarchy
• MORE: AMD vs Intel

AMD Ryzen 7 9800X3D Gaming Benchmarks — The TLDR

The particulars of our test setup are on page four. We used lifted power limits for the Intel processors, which result in higher power consumption and heat, but you get faster performance in exchange.

Nearly every title in our 13-game test suite benefits tremendously from 3D V-Cache, but this isn't the case with all games in the wild. Some games simply don't benefit as much, or at all, from the 3D V-Cache. This can be due to either the game code or a GPU limitation, though the higher 1% lows we measured throughout our suite will be very beneficial even when you are GPU-limited.

Below is the high-level view of gaming performance, using the geometric mean of our 13 gaming benchmarks at 1080p, which is then followed by the individual benchmarks further down the page. We're testing with an Nvidia GeForce RTX 4090 to reduce GPU-imposed bottlenecks as much as possible, and differences between test subjects will shrink with lesser cards or higher resolutions and fidelity.

Keep in mind that the RTX 5090 and RTX 5080 will likely arrive early next year. While 1080p gaming might seem a bit of a waste on the 4090, if the 5090 proves substantially faster — as we suspect it will — the gaps we're seeing here at 1080p would then extend into the 1440p and even 4K range. If you're after maximum gaming performance, the 9800X3D chip will almost certainly remain the best option for the coming year.

We see right away that the $480 Ryzen 7 9800X3D delivers almost unbelievable gaming performance, beating Intel's fastest gaming chip, the $440 Core i9-14900K, by an incredible 30%.

While the 14900K is Intel's fastest gaming chip, it isn't the newest flagship — that falls to the lackluster $620 Core Ultra 9 285K, which hasn't managed to match, let alone exceed, Intel's previous-gen flagship. That means the 285K fares even worse in our overall measurement: AMD claims to beat the 285K by 20%, but in our tests, the 9800X3D is an incredible 35% faster. AMD says the 9800X3D is 8% faster than the $450 Ryzen 7 7800X3D, but again, we measured a much higher delta, with the 9800X3D being 14.7% faster in our test suite.

Looking at the chart of average clock rates during the 13-game test suite is insightful. The 9800X3D ran at an average of 5.2 GHz, which is ~10% faster than the 7800X3D's clock speed. The advantages of Zen 5's higher IPC also weigh in here, helping to explain the impressive gen-on-gen improvements we see in our admittedly cache-friendly test suite.

There are other important factors, too. The stock Ryzen 7 9800X3D's 1% low frame rates (a good smoothness indicator) points to an exceptionally smooth gaming experience. In fact, the 9800X3D's 1% lows weigh in at 148 fps, which is 'only' 2.5% slower than the Core Ultra 9 285K's average frame rate. Impressive indeed.

We also tested with the base stock DDR5-5600 memory configuration for the 9800X3D, marked with '5600' in the above chart, and used DDR5-6000 memory for the standard 'Ryzen 7 9800X3D' listing. Moving up to the faster memory yields a 1.8% speedup on average. (We left that entry out of the individual tests below to make the charts easier to read.) This is a great result with extremely cheap RAM and an entirely different result than shelling out for super fast CUDIMM memory for the Core Ultra chips and receiving a mere 2% improvement.

We also tested with the auto-overclocking Precision Boost Overdrive (PBO) feature enabled (marked with 'PBO' in the chart), and engaging this feature gave us around 2.6% more performance. Note that this will vary based on cooling, chip quality, and other factors. Also, overclocking the competing processors would yield at least some additional performance, too.

The power consumption chart shows that the Ryzen 9 7900X3D averaged 76.6W of power consumption across our suite, while Intel's Arrow Lake ran in the 65 to 67W range, and the previous-gen Ryzen 7 7800X3D consumed 66.4W. However, flipping through to the CPU efficiency chart, which plots FPS-per-watt, finds the 9800X3D and 7800X3D essentially tied for the best efficiency in our test pool, while the Arrow Lake chips are much less efficient overall thanks to their lower gaming performance, falling to the middle of the pack.

Pricing is a big factor, too. Our FPS-per-dollar chart shows that the Ryzen 7 57000X3D is the unquestioned value champ, but the Ryzen 7 9800X3D beats all Intel chips except the Core i7-14700K. Given that the Ryzen 7 9800X3D provides the fastest gaming performance overall, AMD will have absolutely no problem charging a premium.

Our overall measurement of gaming performance is much higher than AMD's for multiple reasons. AMD used multiple games that did not benefit from 3D V-Cache in its tests (our suite is very heavily skewed to 3D V-Cache-friendly titles), it used Windows 23H2 (we use 24H2), and it tested with the performance-sapping Virtualization Based Security (VBS) setting enabled, whereas we test with this feature disabled. We have included AMD's benchmarks with a very broad range of titles at the end of this page to provide a more balanced view of the impact of 3D V-Cache (a few games have 1 to 2% lower performance than with the previous-gen 7800X3D, but that is very limited).

The competition between AMD and Intel chips can vary based on the title (particularly with X3D models) and the GPU you use. It's best to make an informed decision based on the types of titles you frequently play, so be sure to check out the individual game tests below.

A Plague Tale: Requiem Benchmarks — AMD Ryzen 7 9800X3D

Our first gaming chart might look like an abnormally large lead for the 9800X3D, but for 3D V-Cache chips, this is actually pretty much par for the course. Looking just at our 'normal' 9800X3D result (i.e. not the PBO or DDR5-5600 scores), it beats the 7800X3D by 16.4%, only slightly more than our overall geomean. The Intel chips are a different matter, with the 9800X3D showing a 42% lead over the 14900K and a whopping 60% lead over the 285K — yikes!

Baldur's Gate 3 Benchmarks — AMD Ryzen 7 9800X3D

Baldur's Gate 3 really likes X3D as well, and the Zen 5 architecture. This time the 9800X3D beats its predecessor by 25%, with a 44% lead over the 14900K and a 47% lead over the 285K CUDIMM. So this is also one of the few games where Intel's Arrow Lake nearly matches the prior generation Raptor Lake Refresh.

Borderlands 3 Benchmarks — AMD Ryzen 7 9800X3D

It's worth noting that Borderlands 3 was an AMD promotional title, and that might play a role in its performance. But here the 9800X3D only beats the 7800X3D by 11%, with a 30% lead over the 14900K and a 56% lead over the 285K. That's still a huge loss for Arrow Lake, but somewhat closer on the others.

Cyberpunk 2077 Benchmarks — AMD Ryzen 7 9800X3D

Cyberpunk 2077 can be an incredibly GPU heavy game with ray tracing, and we're opting for the more CPU constrained ultra (non-RT) settings, without upscaling — and note that none of our CPU gaming tests have upscaling of any form enabled. Even without RT, it's still more GPU limited than many of our other results, with the 9800X3D delivering a 10% lead over its predecessor, 22% lead over the 14900K, and a still very large 51% lead over the 285K.

F1 2024 Benchmarks — AMD Ryzen 7 9800X3D

We test F1 24 with the ultra preset, which does enable some RT effects. They're relatively lightweight compared to other games, and of course we're still hitting 200+ FPS on most of the CPUs. Here the 9800X3D leads the 7800X3D by just 2%, it's a virtual tie with the 14900K (1% lead to AMD), and 'only' 13% faster than the 285K. This is the closest result we have from our gaming suite, no doubt due to F1 24's default enablement of RT at the ultra preset.

Far Cry 6 Benchmarks — AMD Ryzen 7 9800X3D

Far Cry 6 is another AMD-promoted game, which may impact the standing, though it's by no means one of the biggest outliers for our suite. The 9800X3D beats the incumbent 7800X3D by 14%, right about on par with our average result. But Intel CPUs fall well behind, with a 40% lead over the 14900K and a 43% lead over the 285K. So again, this is a game where Arrow Lake at least wasn't too far behind Raptor Lake.

Final Fantasy XIV Benchmarks — AMD Ryzen 7 9800X3D

Final Fantasy XIV — not the newer Final Fantasy XVI — gives the 9800X3D yet another 16% win over its predecessor. Intel CPUs fail to impress once again, with the new Zen 5 X3D chip beating the 14900K by 39%, and the 285K by 63%. That's the biggest lead over Arrow Lake, if you're wondering.

Hitman 3 Benchmarks — AMD Ryzen 7 9800X3D

Hitman 3 leverages Intel's E-cores for certain game functions, but it also likes Ryzen X3D's voluminous L3 cache. The 9800X3D leads the 7800X3D by 13%, with a less dramatic 15% lead over the 14900K and 'only' a 19% lead over the 285K.

Hogwarts Legacy Benchmarks — AMD Ryzen 7 9800X3D

Hogwarts Legacy is the second most favorable result for 9800X3D compared to 7800X3D, with a 21% overall victory by the new chip. Intel's results are basically in line with our overall average: the 9800X3D ends up 33% faster than the 14900K and 37% faster than the 285K.

Microsoft Flight Simulator Benchmarks — AMD Ryzen 7 9800X3D

Microsoft Flight Simulator 2021 obviously benefits from L3 cache — the Ryzen X3D chips are incredible in this title. 9800X3D beats the 7800X3D by 16%, close to our geomean difference, but Intel's chips fall off the pace. AMD's latest is 38% faster than the 14900K and 42% faster than the 285K. That's not Intel's worst result, but it's definitely in the lower half of the test suite as far as performance goes.

Spider-Man: Remastered Benchmarks — AMD Ryzen 7 9800X3D

Spider-Man: Remastered is a curious result, in that it's one of the few games where Arrow Lake CPUs actually outperform their Raptor Lake predecessors. The 9800X3D ends up just 10% faster than the 285K, our closest test result for those two chips. Against the 7800X3D, it's still a larger than average 17.6% lead, while the 14900K seriously underperforms and the 9800X3D claims a 47% margin of victory. This is the 14900K's worst result overall, though it's only a few percent worse than the Baldur's Gate 3 and A Plague Tale: Requiem standings.

Starfield Benchmarks — AMD Ryzen 7 9800X3D

Starfield is another game that doesn't hate Arrow Lake, though the 9800X3D still wins by 13%. It's also 17.5% faster than the 7800X3D, and 21% faster than the 14900K.

Watch Dogs Legion Benchmarks — AMD Ryzen 7 9800X3D

And we wrap up our gaming results with Watch Dogs Legion, where the 9800X3D lead over the 7800X3D sits at 12.6% — slightly below the geomean. AMD beats the 14900K by 22%, and the 285K by 16.4%, so this is one of the few other results where Arrow Lake comes out ahead of the 14900K.

AMD's Ryzen 7 9800X3D Gaming Benchmarks

As mentioned, performance can vary based on the title. AMD shared its own benchmarks to highlight what you should expect in various titles. As with all vendor-provided performance data, view it with the appropriate skepticism. AMD's test notes are on the final slide.

• MORE: Best CPUs for Gaming
• MORE: CPU Benchmark Hierarchy
• MORE: AMD vs Intel

AMD Ryzen 7 9800X3D Productivity Benchmarks — The TLDR

We boil down productivity application performance into two broad categories: single-threaded and multi-threaded. These slides show the geometric mean of performance in several of our most important tests in each of these categories, but be sure to look at the expanded results below for more granular analysis.

AMD's 3D V-Cache chips have long delivered exceptional gaming performance but lagged in productivity workloads due to the lower clock rates borne of the thermal issues associated with the L3 chiplet. Reducing that disparity was one of the primary motivators for the new design, and here we can see the payoff.

The 8-core Ryzen 7 9800X3D is 4% faster in multi-threaded workloads than the 8-core Ryzen 7 9700X, which leverages a traditional CPU design. That's quite an accomplishment, and it demonstrates that AMD is putting the extra 15W of TDP headroom over the 9700X (tested in 105W mode) to exceptionally good use. We also see a big 34% increase in multi-threaded performance over the prior-gen 7800X3D.

The competing Intel chips are much faster in threaded work, though. The Core i9-14900K is 41% faster in threaded workloads, while the Core Ultra 9 285K is a whopping 64% faster. Keep in mind that the 14900K currently costs less than the 9800X3D.

The Ryzen 7 9700X has a 300 MHz higher boost clock rate than the 5.2 GHz Ryzen 7 9800X3D, so it isn't surprising to see the former take a ~3% lead in single-threaded work, a trend that you'll see repeated throughout the benchmarks below. While the Ryzen 7 9800X3D has made big generational improvements — it's 14% faster than the 7800X3D in single-threaded work — it still lags the competing Intel chips by significant margins. The Core Ultra 9 285K is 13% faster in single-threaded work, while the Core i9-14900K is 10% faster.

The Ryzen 7 9800X3D delivers truly exceptional gaming performance, but it's important to remember that similarly-priced processors provide much more performance in productivity work, so you should buy based on your priorities. You'll notice that the Ryzen 5 5700X3D, which delivers absolutely phenomenal performance for its price point in gaming, suffers tremendously in the application benchmarks due to its aging Zen 3 architecture, so remember to proceed with caution when selecting these chips over more well-rounded competitors. They truly are specialist chips for gaming. 

Rendering Benchmarks — AMD Ryzen 7 9800X3D

Intel has made strong gains with Arrow Lake, as we can see throughout this series of benchmarks. The increased throughput from the CUDIMMs doesn't often make a big impact, so unless there's a pricing benefit or future firmware updates that change the situation, regular DDR5 DIMMS seem just fine.

We see some remarkable gains in threaded workloads for the 285K, with performance in Cinebench 2024, POV-Ray, C-Ray, and Luxmark all taking giant strides forward. However, the AMD Ryzen 9 9950X remains strong in several tasks, like Embree, V-Ray 6, Blender, and Corona.

Encoding Benchmarks — AMD Ryzen 7 9800X3D

Most encoders tend to be either heavily threaded or almost exclusively single-threaded — and it takes an agile chip to master both disciplines. Handbrake, SVT-HEVC, and SVT-AV1 serve as our threaded encoders, while LAME and FLAC are indicative of how the chips handle lightly-threaded engines.

Adobe Photoshop, Premiere Pro, DaVinci Resolve — AMD Ryzen 7 9800X3D

The PugetBench for Creators suite is a wonderful tool for benchmarking multiple types of Adobe applications, not to mention DaVinci Resolve. We used the benchmark for Photoshop, Premiere Pro, and DaVinci workloads.

Web Browser, Office Benchmarks — AMD Ryzen 7 9800X3D

Compilation, Compression, AVX, Chess Engines, Others — AMD Ryzen 7 9800X3D

This selection of tests runs the gamut from massively parallel molecular dynamics simulation code in NAMD to compression/decompression performance.

Y-cruncher computes Pi with the AVX instruction set, making for an exceedingly demanding benchmark. This benchmark was recently updated with specific tuning for AMD's AVX-512 implementation. The new code delivers a big boost to Ryzen. Intel's new chips make big strides in this benchmark over the prior generation, but they only support AVX-256 natively, ultimately giving AMD an easy win.

Intel has historically performed well in compilation tasks, and here we can see the 285K taking back the leadership spot in the LLVM benchmark, albeit by a slim margin. The NAMD simulation code benchmark finds the 285K making a much more definitive claim for leadership. 

Geekbench 6, Geekbench AI — AMD Ryzen 7 9800X3D

The GeekBench AI results in the slide deck above show performance with the workload running on the CPU cores via OpenVino and ONNX. Only two CPUs in this group have in-built NPUs — the Core Ultra 9 285K and Ultra 7 265K — making them the only processors capable of running the workload on a dedicated NPU. Given that the NPU is focused on low-power inference tasks, the results are impressive relative to the performance when the workload executes on the CPU cores. 

SPECworkstation 3.1 — AMD Ryzen 7 9800X3D

The SPECworkstation 3 benchmark suite is designed to measure workstation performance in professional applications. The full suite consists of more than 30 applications split among seven categories, but we've narrowed the list to tests that largely focus on CPU performance. We haven't submitted these benchmarks to the SPEC organization, so be aware that these are not official benchmarks.

• MORE: Best CPUs for Gaming
• MORE: CPU Benchmark Hierarchy
• MORE: AMD vs Intel

AMD Ryzen 7 9800X3D Power Consumption and Efficiency

Here we can see that the Ryzen 7 9800X3D doesn't draw as much power as the Intel comparables, thus resulting in a cooler and quieter system with more forgiving PSU, motherboard, and cooling requirements.

Note that AMD's chips generally have a higher idle power draw. I've measured this on a few motherboards, so this might be applicable across the AM5 ecosystem. AMD's AM5 chips can draw up to ~35W of power during idle compared to 12W with Intel's Core Ultra, a significant disadvantage. This is certainly something AMD should look to address with future firmware updates, as there's little reason for a modern CPU architecture to idle at such high power levels.

The final two charts take a slightly different look at power consumption by plotting the cumulative energy required for an x265 HandBrake and Blender workload. We plot this 'task energy' value in Kilojoules on the left side of the chart, with performance charted on the x-axis.

These workloads consist of a fixed amount of work, so we can plot the task energy against the performance during the job (bottom axis), thus generating a useful power chart. Faster compute performance and lower task energy are ideal. That means processors closest to the bottom right corner of the chart are the best.

Test Setup

All testing was conducted with Microsoft Windows 24H2, and all drivers and applications were updated to the latest versions. To speed up the testing process, we use XMP/EXPO as the default memory profiles for all tested configurations. The memory speeds used for each chip are shown in the table below.

We generated all the gaming results for the Ryzen 9000 processors with the new 1.2.0.2a AGESA. Due to time constraints, we haven't yet fully retested all of the Ryzen 9000 processors in our productivity test suite with the new 1.2.0.2a AGESA — for all but the Ryzen 7 9800X3D, our productivity benchmarks are with the 1.2.0.2 revision.

Microsoft has advised gamers to turn off several security features to boost gaming performance. For maximum performance, we disabled Virtualization Based Security (VBS) on all systems. Be aware that, due to hardware acceleration, some processor architectures handle virtualization better than others. This can provide a performance advantage in gaming with VBS enabled. We're working to quantify the performance differences and may adjust our VBS policy in the future. The table below provides further hardware details.

• MORE: Best CPUs for Gaming
• MORE: CPU Benchmark Hierarchy
• MORE: AMD vs Intel

The AMD Ryzen 7 9800X3D is the fastest gaming chip on the market by a large, almost unbelievable margin, easily beating Intel's competing processors. Yes, the $480 price tag is expensive, leaving room for like-priced alternatives with a more balanced performance profile in productivity applications. However, many of those competing chips come from within AMD's own stable — Ryzen 9000 has improved quite a bit with the recent firmware and Windows updates — and the fact is that AMD offers the fastest gaming performance in the world, so it can charge for it.

AMD has significantly improved performance in productivity workloads, too, helping to eliminate some of the trade-offs of selecting a gaming-optimized X3D chip. It has also fully opened up the overclocking bells and whistles, giving enthusiasts more options for tuning. You should keep your expectations in check for big gains from PBO, but this simple largely-automated feature does help in some applications.

AMD will have no problem selling the Ryzen 9 9800X3D at its price point, and Intel doesn't appear to have a chip that can compete in gaming in the near future, once again ceding the gaming crown to AMD for the indeterminate future.

Below, we have the geometric mean of our gaming test suite at 1080p and a cumulative measure of performance in single- and multi-threaded applications. These are cumulative measurements, so outliers in specific titles can impact the values. The Ryzen X3D chips use 3D V-Cache tech, which doesn’t accelerate all titles equally, and the Zen 3 and 4 versions come with performance penalties in productivity applications. Be sure to check our individual game and productivity benchmarks in the preceding sections to make an informed decision based on your intended workloads.

We conducted our gaming tests with an Nvidia RTX 4090, so performance deltas will shrink with lesser cards and higher resolution and fidelity settings. And conversely, when the RTX 5090 comes out, deltas could actually increase — we'll likely find out how much early next year.

As we said earlier, the competition isn't even close — the $480 Ryzen 7 9800X3D delivers outstanding gaming performance, beating Intel's fastest gaming chip, the $440 Core i9-14900K, by 30% in our test suite. Meanwhile, the 9800X3D is 35% faster than the current-gen Intel flagship, the $620 Core Ultra 9 285K, which hasn't managed to match, let alone exceed, Intel's previous-gen flagship. The stock Ryzen 7 9800X3D's 1% low frame rates (a good smoothness indicator) also deliver an exceptionally smooth gaming experience, which will benefit gamers even in GPU-limited scenarios.

The Core i9-14900K is a closer competitor for the 98000X3D due to its lower $440 price tag, and it does provide better performance in productivity workloads at its price point than the 9800X3D, beating it by 41% and 10% in multi- and single-threaded workloads, respectively. If you’re more interested in productivity work in this price range, the 14900K is a solid contender, albeit one that typically draws more power (except at idle).

Another benefit for AMD's CPU is that you won’t have to kit the 9800X3D out with expensive supporting componentry to extract the best of 3D V-Cache. The chip is relatively tame in terms of power and thermal requirements, exhibiting great power and efficiency metrics in our tests, and even cheap DDR5-5600 memory kits deliver within a few percentage points of performance of faster DDR5-6000 kits — the X3D chips simply don’t need EXPO memory to run at near-peak performance. That stands in stark contrast to splurging on a pricey CUDIMM kit for Arrow Lake and only scraping out ~2% more gaming performance.

The Ryzen 7 9800X3D does have a $30 higher introductory price tag than its predecessor, but you get what you pay for. Our FPS-per-dollar chart shows the Ryzen 7 9800X3D beats all Intel chips except the Core i7-14700K in terms of value. If you’re scraping together a system on a budget, you can always opt for a prior-gen 600-series motherboard, too.

Given that the Ryzen 7 9800X3D provides the fastest gaming performance overall, we expect that this chip will find no shortage of enthusiasts ready to fork over their hard-earned dollars. We also expect AMD to launch the larger dual-CCD X3D models in the coming months, so it might be best to hold off if you’re looking for something with a bit more threaded horsepower. With the move to put the 3D V-Cache die under the processing core, we could even see something like a 9950X3D with double the stacked cache — and if that happens, it might end up being the no-holds-barred performance champion for the current generation, regardless of workload.

But for now, if you want the best gaming chip on the market, there simply is no better option than the Ryzen 7 9800X3D.

• MORE: Best CPUs for Gaming
• MORE: CPU Benchmark Hierarchy
• MORE: AMD vs Intel

The Corsair MP700 Elite delivers on the possibilities promised in our Phison E31T preview. It's one of the first of what will surely be many retail products using the controller and NAND combination. Most importantly, it's here now and you can hold it in your hands — or put it into a PC where it belongs. The Corsair MP700 Elite comes from a combination of efforts, with Phison delivering a production 7nm 'N7' SSD controller and Kioxia via its high-speed BiCS8 flash. The result is a powerful, efficient, and optimistic sign of things to come in the storage space.

The MP700 Elite is an exciting product not only because it treads new ground, but because it does it simply and effectively. Corsair is sticking to the most popular capacities and keeping the heatsink as an option. It's a continuation of the company’s SSD strategy of focusing on what sells well, combined with getting new technology out to the public as early as possible. There's room for some improvements with the follow-through, particularly in getting the website updated, getting stock to resellers more quickly, and setting drive prices at a reasonable place in a more timely fashion. But Corsair has generally nailed getting the right products out there, and it just needs to do that more smoothly.

We'll see many more drives like the MP700 Elite from competitors, including the contemporaneous Addlink G55 that's also on our testbench. Corsair comes with a reputable brand name and a willingness to stay relevant in the SSD market. We think this technology is the future as it combines high performance with high power efficiency in a svelte, single-sided design. There’s little not to like about the MP700 Elite, although it doesn’t quite fulfill all storage roles and we’d like to see a 4TB SKU.

 Corsair MP700 Elite  Specifications

Corsair focused on the most popular capacities with the MP700 Elite launch, only offering the drive at 1TB and 2TB capacities — at least for now. There have been increasing calls for 4TB drives in recent months from storage enthusiasts, and those make the most sense for high-end products like the Samsung 990 Pro. Of course, the MP700 Elite isn't remotely what we'd classify as a budget drive and in fact qualifies as a high-end offering, but current flash isn't dense enough to make 4TB optimal for the four-channel SSD controller used here, not that it’s prevented such drives from existing and even being popular.

The market right now still has 2TB as the sweet spot, and Corsair is focusing on delivering that with a cutting-edge drive like the MP700 Elite. Corsair didn’t provide official pricing for this drive prior to launch, but it’s already showing up at some vendors with prices of $149.99 for 1TB and $249.99 for 2TB. Those prices feel absurdly high even for a drive of this quality, and we expect street prices will be significantly lower considering they aren’t too far off the MP700 Pro — which costs $279 for 2TB and $165 for 1TB at present.

The drive’s specifications closely match the Phison E31T engineering sample we previously previewed. The MP700 Elite can hit up to 10,000 / 8,500 MB/s for sequential reads and writes and up to 1,300K / 1,400K random read and write IOPS. These performance targets are identical for both 1TB and 2TB models, which means you’re not giving anything beyond capacity up by sticking with the smaller SKU. The warranty is standard at five years with up to 600TB of writes per TB capacity.

Corsair MP700 Elite Software and Accessories

The primary download for Corsair SSDs is the Corsair SSD Toolbox. It provides a single application with a variety of functionality, allowing you to verify drive health, search for firmware updates, secure erase the drive, backup data, clone a drive, and more. The quality of the toolbox varies from manufacturer to manufacturer — if they even offer one — with Corsair’s being one of the better options. Corsair’s been in the SSD business for quite a while, so its name is more well-known and having an SSD toolbox helps reinforce that brand. If software and updates are important to you, this is a selling point.

Corsair MP700 Elite: A Closer Look

The MP700 Elite is, unsurprisingly, single-sided. This makes it ideal for laptops and other devices that only accept single-sided drives, and it can also make the drive easier to cool. Nevertheless, Corsair offers a heatsink version of the drive if that’s your thing. The heatsink is aluminum and increases the width from 22mm to 24mm and height or thickness from 2mm to 9mm, so make sure your system can handle that if you go that route.

The heatsink should be extremely effective for a drive like this, one that already runs generally cool. Even though the label suggests up to 7.26W of power draw — 3.3V and 2.2A — the spec sheet pegs this drive at less than 6W at peak, which is about half that of the high-end PCIe 5.0 SSDs.

Peeling back the label reveals Kioxia 218-Layer BiCS8 TLC flash, the same we found on the 2TB E31T ES. Each NAND package offers 1TB of capacity, made up of eight 1Tb dies. Potentially the packages could be arranged to allow four per side, but the current layout would only allow for a 4TB single-sided drive with sixteen dies per package.

The controller is also the same E31T (the "T" suffix is for DRAM-less Phison controllers), this one manufactured during week 33 of this year. We discussed this hardware in some depth in our ES preview. If you have a drive with the Phison E21T or E27T controller, you more or less know what you’re getting here. The difference is that this one is manufactured using TSMC's N7 process node, which allows it to push higher speeds with lower power usage. And of course, this controller supports PCIe 5.0 where those others are PCIe 4.0 models.

MORE: Best SSDs

MORE: How We Test HDDs And SSDs

MORE: All SSD Content

Comparison Products

The Corsair MP700 Elite is one of the first retail drives based on the E31T controller, although we have the Addlink G55 waiting in the wings. (Spoiler alert: It's effectively the same performance, just without the Corsair brand name.) The MP700 Elite isn't intended to hang with drives built on the faster Phison E26 controller, or at least not the higher-end variants that came after the initial launch of 10 GB/s PCIe 5.0 SSDs. The Crucial T705 shows what that latter controller can do, with no comparable drive available with SMI’s SM2508 just yet but with the Adata Legend 970 Pro available using InnoGrit’s competing IG5666 controller.

Drives that might be in more danger include the most popular, high-end PCIe 4.0 SSDs. These consist of the Samsung 990 Pro, the WD Black SN850X, and the SK hynix Platinum P41. It may be more fair to compare the MP700 Elite with the four-channel Crucial T500, but remember that it and the previous three drives listed all have DRAM. There’s some question about whether or not DRAM is needed today, but arguably it's still useful for workhorse drives.

Directly comparable drives would be DRAM-less and these include the Samsung 990 EVO Plus and the Silicon Power US75, the latter representing a host of popular drives including the Teamgroup MP44, the Lexar NM790, and the Addlink A93. At least one of these drives, the MP44, has been found with the E27T and BiCS6 flash instead of the launch MAP1602 and 232-Layer YMTC TLC flash, for which Corsair also has an answer with its Corsair MP600 Elite. It’s also possible to use QLC flash, as is the case with the Crucial P310, but such drives should be less expensive with a heavier focus on capacity. 

Trace Testing — 3DMark Storage Benchmark

Built for gamers, 3DMark’s Storage Benchmark focuses on real-world gaming performance. Each round in this benchmark stresses storage based on gaming activities including loading games, saving progress, installing game files, and recording gameplay video streams. Future gaming benchmarks will be DirectStorage-inclusive and we also include notes about which drives may be future-proofed. 

The MP700 Elite easily matches or beats any drive that came before it, excepting the high-end PCIe 5.0 T705 and similar drives using the E26 controller. The T500 deserves an honorable mention, demonstrating that newer flash and DRAM can help a drive stay relevant. Both the T500 and MP700 Elite have Phison’s DirectStorage-optimized I/O+ firmware which may become more important with future game titles.

Trace Testing — PCMark 10 Storage Benchmark

PCMark 10 is a trace-based benchmark that uses a wide-ranging set of real-world traces from popular applications and everyday tasks to measure the performance of storage devices. The results are particularly useful when analyzing drives for their use as primary/boot storage devices and in work environments.  

PCMark 10 performance is also exceptional. If you really want the most bandwidth and best application performance, an E26-based drive like the T705 or Sabrent Rocket 5 is still the best option. There’s no substitute for 14+ GB/s of bandwidth with a DRAM-equipped, eight-channel controller, but that’s not to say the MP700 Elite is a slouch — it surpasses the best PCIE 4.0 drives in this test.

Console Testing — PlayStation 5 Transfers

The PlayStation 5 is capable of taking one additional PCIe 4.0 or faster SSD for extra game storage. While any 4.0 drive will technically work, Sony recommends drives that can deliver at least 5,500 MB/s of sequential read bandwidth for optimal performance. In our testing, PCIe 5.0 SSDs don’t bring much to the table and generally shouldn’t be used in the PS5, especially as they may require additional cooling. Check our Best PS5 SSDs article for more information.

Our testing utilizes the PS5’s internal storage test and manual read/write tests with over 192GB of data both from and to the internal storage. Throttling is prevented where possible to see how each drive operates under ideal conditions. While game load times should not deviate much from drive to drive, our results can indicate which drives may be more responsive in long-term use.

We’ve adjusted our PS5 scores so everything lines up and, unsurprisingly, the MP700 Elite is right in the mix. It’s only running at PCIe 4.0 speeds in the PS5, which limits its potential to some degree, but it will be very efficient in this mode. We wouldn’t recommend using it for a PS5 unless the price is right, as there are plenty of 4.0 drives that can provide an equivalent experience.

Transfer Rates — DiskBench

We use the DiskBench storage benchmarking tool to test file transfer performance with a custom, 50GB dataset. We write 31,227 files of various types, such as pictures, PDFs, and videos to the test drive, then make a copy of that data to a new folder, and follow up with a reading test of a newly-written 6.5GB zip file. This is a real world type workload that fits into the cache of most drives.

The MP700 Elite’s PCIe 5.0 interface allows it to have a higher throughput than 4.0 drives. However, it can’t match a true monster like the T705. You do at least get faster transfers with this drive over high-end PCIe 4.0 options. Our DiskBench test is only 50GB in size so it shows a typical real-world example, while our write saturation test found below is better for gauging sustained writes, and it also indicates how larger writes on a fuller drive might run.

Synthetic Testing — ATTO / CrystalDiskMark

ATTO and CrystalDiskMark (CDM) are free and easy-to-use storage benchmarking tools that SSD vendors commonly use to assign performance specifications to their products. Both of these tools give us insight into how each device handles different file sizes and at different queue depths for both sequential and random workloads. 

Our ATTO results for the MP700 Elite are somewhat more consistent than with the original E31T ES, but it does appear that more optimization is possible. In part this might be due to the flash being used, which is novel, considering that Phison’s other most recent controllers have leaned more heavily on Micron flash. You can see shades of this with CDM’s sequential read QD1 result, where the T500 actually comes out ahead. With queue depth and with writes, however, the MP700 Elite has no problem running away from the PCIe 4.0 SSD field.

More impressive is the small I/O, low queue depth performance. Looking at random 4KB latencies, both for reads and writes, the MP700 Elite delivers. We discussed possible reasons for this in our E31T ES review but, to quickly recap, the faster I/O speed and the architecture of the flash both support improvements in this area. If you’re someone who really focuses on this type of performance, you’re not sacrificing anything with this drive.

Sustained Write Performance and Cache Recovery

Official write specifications are only part of the performance picture. Most SSDs implement a write cache, which is a fast area of pseudo-SLC (single-bit) programmed flash that absorbs incoming data. Sustained write speeds can suffer tremendously once the workload spills outside of the cache and into the "native" TLC (three-bit) or QLC (four-bit) flash. Performance can suffer even more if the drive is forced to fold, which is the process of migrating data out of the cache in order to free up space for further incoming data.

We use Iometer to hammer the SSD with sequential writes for 15 minutes or more (60 minutes in this case) to measure both the size of the write cache and performance after the cache is saturated. We also monitor cache recovery via multiple idle rounds. This process shows the performance of the drive in various states as well as the steady state write performance.

Usually manufacturers use favorable testing conditions for drive specifications, and we cover those results with our CDM/ATTO testing, as well as here with a 1MiB block size at QD32. The MP700 Elite’s pSLC sustained write speed hits up to 8.8 GB/s, higher than the spec sheet’s 8.5 GB/s. It can sustain this for 50 seconds for a ~440GB cache, which is large but far from being as large as it could be.

This helps improve “quality of service” or QoS, which essentially means that the drive remains reasonably responsive even when hit with an expectedly long workload. A larger cache can sustain maximum performance longer, but when a drive is realistically used it’s going to be fuller and that dynamic cache will in turn be smaller. Thus it’s important for the performance state, in the rare case the cache runs out, to be better in order to prevent a jarring transition and a weaker long tail response.

This is realized with the direct-to-TLC performance state, which achieves 1.5 GB/s or so once the pSLC cache runs out on the MP700 Elite. This sounds slow in comparison to the rated write speed but is actually pretty decent for a native flash speed. Faster is certainly possible, but there's always a need to balance the cache size with this state’s performance level. It’s only after a truly extreme write duration that the drive falls to folding at around half this speed, which is still better than anything you see from QLC flash even in its native state.

Power Consumption and Temperature

We use the Quarch HD Programmable Power Module to gain a deeper understanding of power characteristics. Idle power consumption is an important aspect to consider, especially if you're looking for a laptop upgrade as even the best ultrabooks can have mediocre stock storage. Desktops may be more performance-oriented with less support for power-saving features, so we show the worst-case.

Some SSDs can consume watts of power at idle while better-suited ones sip just milliwatts. Average workload power consumption and max consumption are two other aspects of power consumption but performance-per-watt, or efficiency, is more important. A drive might consume more power during any given workload, but accomplishing a task faster allows the drive to drop into an idle state more quickly, ultimately saving energy.

For temperature recording we currently poll the drive’s primary composite sensor during testing with a ~22°C ambient. Our testing is rigorous enough to heat the drive to a realistic ceiling temperature.

It’s difficult to understate the impressiveness of the MP700 Elite’s power efficiency. This is a drive that can hit performance numbers that would have seemed impossible only a few years ago and it does it with a DRAM-less, four-channel controller. This is good news for the future of SSDs, as this is ostensibly a budget-oriented drive — we'll have to see how prices change in the coming months. While in real-world terms the user experience isn’t greatly improved with faster SSDs and the average power draw of any SSD is relatively low, there’s no one who is going to turn down more efficient hardware.

With the MP700 Elite’s lower power draw — being more efficient doesn’t mean the average or max consumption numbers are necessarily lower, as the higher performance level can finish workloads faster — cooling is much less of a concern. This drive can be run without a heatsink, and that makes it great for laptops. On the other hand, in some scenarios you might want a heatsink and Corsair allows for that by offering a heatsinked option. A heatsink can improve aesthetics and, in hot environments or where you might be running heavier workloads, it keeps the drive temperature in check.

Test Bench and Testing Notes

We use an Alder Lake platform with most background applications such as indexing, Windows updates, and anti-virus disabled in the OS to reduce run-to-run variability. Each SSD is prefilled to 50% capacity and tested as a secondary device. Unless noted, we use active cooling for all SSDs. 

Corsair MP700 Elite Bottom Line 

The Corsair MP700 Elite is the quintessential “budget” PCIe 5.0 SSD. It has excellent performance in every test, usually beating every 4.0 drive under the sun. This is an accomplishment for a four-channel, DRAM-less solution all on its own. That it manages this with exceptional power efficiency is the icing on the cake. You can finally put a retail PCIe 5.0 SSD into your laptop! Heck, it’ll be even more efficient in a 4.0 slot. The sustained write performance could be better, sure, but it’s still pretty good for this class of drive and the large cache should handle anything you throw at it.

This high level of power efficiency means you can run the drive without a heatsink, but Corsair still offers that option. This is a nice touch as the MP700 Elite is still a high-performance drive. It will be interesting to see what happens when faster TLC flash comes out and, for that matter, when denser QLC flash is available. It will soon become possible to have fast, efficient, capacious storage without some of the question marks DRAM-less drives used to bring and without the compromises crazy high-end drives require. That said, the Crucial T500 provides a good counter-argument with 4TB of fast TLC flash and DRAM to boot.

If we have an issue with the MP700 Elite, it’s with the pricing. We need to see this drive on the market for a while, until its availability and pricing settle down. It could be priced at the same level as high-end PCIe 4.0 solutions and it would be tough to have a problem with that, given its performance results. The problem there is that the T500 can butt its way in and undercut that price very effectively, especially as it has a 4TB SKU. The MP700 Elite would still be the choice for PCIe 5.0 laptops, but it could be a harder sell in other areas. True storage enthusiasts can always jump up to the E26-based drives in a desktop and, for the PS5, the MP700 Elite is just as much overkill as other PCIe 5.0 drives.

That leaves us impressed but patient. This is new technology that pushes the bar higher, but it still needs to be practically priced in a market that’s quite volatile. On paper and in testing it hits all the right notes to be the “perfect” all-around drive, even in a PCIe 4.0 slot, but it could be better with some optimization and better pricing — hopefully, through competition. In that way it’s still a drive of the future rather than of now, but we have no problem recommending it if you want that kind of experience today.

If you’re looking for 4TB, then drives in the realm of the Teamgroup MP44 remain most viable right now, and these drives are simply more affordable at lower capacities, too. If you need DRAM there’s the T500, and if you need higher sustained performance there are the tried-and-true high-end PCIe 4.0 options. The MP700 Elite’s overall performance is still quite excellent against any of these drives, though, and if you want the best power efficiency — and especially if you want that in a 5.0 drive — the MP700 Elite should be the first on your list. 

MORE: Best SSDs

MORE: How We Test HDDs And SSDs

MORE: All SSD Content