AMD's Ryzen AI Max+ 395, popularly known as Strix Halo, has earned a reputation as the scrappy AI enthusiast platform of choice in the year since its introduction, thanks to its powerful integrated Radeon 8060S graphics and support for RAM pools of up to 128GB.

Corsair's AI Workstation 300 brings Strix Halo to the company’s lineup for the first time. With configuration options ranging up to the flagship Ryzen AI Max+ 395 with 128GB of RAM and two 2TB NVMe SSDs in an aluminum shell that takes up just XX liters, this system is a compact and classy implementation of AMD’s flagship APU.

And thanks to its 16-core, 32-thread Zen 5 x86 CPU, the AI Workstation 300 can offer native gaming support under both Linux and Windows that competitors like Nvidia’s DGX Spark and Apple’s various Macs lack.

Corsair provides plenty of connectivity options for general PC usage, unlike the more focused DGX Spark. The front panel boasts an SD card reader, a USB 4 Type-C port, two USB 3.2 Gen 2 Type-A ports, and a combo audio jack.

The rear panel has a DisplayPort 1.4 connector, an HDMI 2.1 output, another USB 4.0 Type-C port with DisplayPort Alt Mode support, a USB 3.2 Gen 2 Type-A port, two USB 2.0 Type A ports, a 2.5 Gigabit Ethernet jack, and another audio combo jack.

Inside, expansion options are limited due to the highly integrated nature of Strix Halo.

It’s easy enough to remove all of the screws at the back of the unit and slide the integrated motherboard and cooling assembly out of the front of the chassis, but the only potentially user-accessible expansion options are a pair of M.2 2280 slots, both of which are populated with 2TB NVMe SSDs in our review system’s configuration.

The AI Workstation 300 comes with Windows 11 out of the box, but its dual-SSD config makes it easy to install and dual-boot Linux if Windows isn't your preferred AI development environment, which is quite handy.

Since our recent AI testing relies in part on llama.cpp and ComfyUI, both of which are better documented and whose behavior is better characterized under Linux on AMD platforms, I installed Ubuntu 24.04 LTS on the spare 2TB drive of my AI Workstation 300, and the subsequent process of installing AMD's drivers and the ROCm stack was straightforward enough.

Corsair AI Workstation 300 Specs

A quick look at AI and general desktop performance

Since Corsair positions the AI Workstation 300 as, well, an AI workstation, we put its AI performance to to the test with the llama.cpp model runner for LLMs and ComfyUI for creative workflows under Ubuntu 24.04 LTS, using ROCm 7.1.1 as our compute backend.

We used Nvidia's DGX Spark and its GB10 chip as our primary point of comparison, since its 128GB of RAM, similar memory bandwidth, and compact form factor make it a natural competitor for Strix Halo generally and the AI Workstation 300 specifically.

We bench two phases of LLM interaction: prompt processing or prefill for input, and token generation for output. We offer 256 input tokens to each model and generate 256 output tokens across a range of context lengths to illustrate what might happen as one's interactions with a model grow over time.

We’ll start off our benchmarks using Meta’s llama-3.1-8B, a dense LLM (i.e., one for which all parameters are activated per token) whose behavior is well-understood in the current AI landscape.

We extended our dense model testing using Google’s Gemma 3 12B and 27B.

We’ll also look at performance with OpenAI’s gpt-oss 120b and Qwen3-30B-A3B models, which use mixture-of-experts (MoE) architectures that are closer to the state of the art in LLM research.

MoE models can have massive parameter counts, but only a portion of those parameters are activated for any given token. That allows MoEs to balance both capability and performance, although they can still require large amounts of memory. GPT-OSS 120b, for example, requires about 60 GB of RAM.

Across our LLMs, the AI Workstation 300 is behind in the compute-heavy prompt processing phase of our benchmarks due to the lower raw compute horsepower of the Radeon 8060S compared to GB10.

In the memory-bandwidth-dependent token generation phase of the workload, the AI Workstation 300 can put up a good fight when context lengths are short, but as they grow, Nvidia's GB10 maintains higher and more consistent performance.

LLMs are just one usage for AI compute, of course. Image generation is another popular task, and even local video generation is now a possibility thanks to relatively compact text-to-video models like LTX-2.

We tested the latest Flux.2 Klein 9B image generation workflow in ComfyUI using the same random seed as a base on both the Spark and the AI Workstation 300. After the first load for the workflow, the AI Workstation 300 needed roughly four times as long to generate an image as GB10 did. That, too, is due in part to GB10's higher raw compute capacity versus the Radeon 8060S.

We also gave LTX-2 a try on the AI Workstation 300, but as of the time of our tests, trying to run the same LTX-2 workflow on the ROCm 7.1.1 stack resulted in HIP errors that would hang ComfyUI or crash the entire GNOME desktop environment.

That software immaturity isn’t Corsair’s fault, but it is an unavoidable consequence of building a box around a highly integrated platform like Strix Halo. We’re going to keep an eye on AMD’s continuing efforts to improve ROCm stability and performance, and we’ll see how it changes the appeal of client AI platforms like the AI Workstation 300 and Strix Halo in 2026.

AI isn’t the only reason one might want a local workstation like this, though. The 16 cores and 32 threads of the Ryzen AI Max+ 395 could potentially outrun the 10 performance and 10 efficiency cores of the DGX Spark’s Arm CPU, so we did some supplemental testing with Geekbench 6 under Ubuntu 24.04 to find out.

The Cortex-X925 performance cores and the Zen 5 cores in the Ryzen AI Max+ 395 are closely matched in the single-threaded portion of Geekbench 6. However, the Ryzen AI Max+ 395 pulls 11% ahead in the multi-threaded portion of the benchmark. That means the AI Workstation 300 will be somewhat faster than the DGX Spark in general-purpose parallel computing tasks like code compilation.

For fun, we also fired up the Unigine Superposition benchmark on both platforms under Ubuntu to get a broad idea of gaming performance.

To be clear, if you’re a gamer first and foremost and don’t need massive amounts of system RAM or VRAM for AI, any PC you can build with even a midrange discrete GPU will vastly outperform both Strix Halo and GB10. But if you’re a local AI enthusiast who wants to game on the side, it’s worth knowing what you’re signing up for.

Despite having 2.4 times as many shader ALUs in its GPU as the Radeon 8060S, the DGX Spark turns in a Superposition score just 4% higher than that of the AI Workstation 300. It’s clear from that result that the limited memory bandwidth of Nvidia’s platform strangles gaming performance. The desktop RTX 5070, which has the same 6144 CUDA cores but much higher memory bandwidth, turns in a Superposition score more than twice as high as that of GB10.

In addition to its roughly equivalent performance, the AI Workstation 300’s main advantage for gaming versus the DGX Spark is that it can boot straight into Windows and get right to it with DirectX titles, all without futzing with x86 emulation layers like FEX. That means you shouldn’t run into show-stopping compatibility issues as we did with Black Myth Wukong on the Spark under Linux.

But as with the DGX Spark, the AI Workstation 300 is an extremely costly way to get a merely passable gaming experience if you don’t strictly need 128GB of RAM to start with. Any traditional gaming PC will be much quicker.

Power, cooling, noise, and thermals

As a highly integrated SoC with mobile roots, the Ryzen AI Max+ 395 has configurable TDPs ranging from 45W to 120W. Corsair exposes them through a front-panel button that switches among three firmware power profiles: Balanced (the default), Max, and Quiet, in that order. Balanced appears to give the SoC a power budget of about 85W, Max around 120W, and Quiet, just 55W.

Each press triggers a brief on-screen indication of the change under Windows, but under Linux, you don’t get any warning that anything has changed, and these settings persist across reboots.

If you casually brush this button, you’re likely to end up in the high-performance Max mode, but you could just as easily end up in the low-performance Quiet mode. Unless you’re paying close attention to the noise levels of the system or have it plugged into a power meter at the wall, you might not notice anything amiss.

It would have been nice if Corsair had put an LED indicator or something on the front of this system to show the active power mode regardless of the OS in use.

In all three power modes, because the system power budget is shared between CPU and GPU, fully loading up both processors at once will reduce performance across both functional units. It’s unlikely that you’ll ever stress the CPU and GPU that way in regular use, but it’s worth understanding that the AI Workstation’s shared power and thermal budget won’t deliver peak CPU and GPU performance with that kind of all-out workload if you do happen to have one.

The AI Workstation 300’s chassis and thermal system is more than up to the task of cooling Strix Halo across all of those power modes. Two blower-style fans cool the SoC by moving air through a fin stack that’s connected to the SoC with three heat pipes, and another slim fan draws in air on the other side of the motherboard for the twin NVMe SSDs. The 300W Flex ATX PSU has a tiny fan of its own.

While we appreciate its thorough ventilation, the AI Workstation 300 isn’t the quietest mini-PC around. It’s silent when at 100% idle, but it audibly spins up its fans under light load, and its noise signature is more prominent than the much smaller DGX Spark when either the CPU or GPU are working hard, especially in Max mode.

We measured peak noise levels of just 39 dBA from 18”/46 cm away in the AI Workstation 300’s Balanced mode during ComfyUI image generation, compared to 37.5 dBA from the DGX Spark under the same workload. That rises to 43.4 dBA in Max mode under the same workload, although the extra juice and noise barely makes a difference for time to completion in ComfyUI.

Warranty and support

Since so many Strix Halo boxes feature largely similar feature sets, the post-purchase experience is one way for a company to stand out. Corsair’s warranty and service are a cut above the competition. The company offers a two-year warranty on the AI Workstation 300 with advance replacement options available.

Most lower-cost Strix Halo boxes on Amazon only offer a one-year warranty, and there's no guarantee that you'll be dealing with timely or local tech support in the United States.

Bottom line

As Strix Halo boxes go, Corsair's AI Workstation 300 is a solid one. It's a sleek, well-integrated package with effective cooling that delivers all of the performance potential of the Ryzen AI Max+ 395. Its classy black aluminum shell and its integrated power supply make it more elegant and portable than other ultra-tiny systems in this vein that need an external power brick that can be lost or forgotten on the move.

This system’s 16-core, 32-thread Zen 5 x86 CPU is great for both single-threaded and multithreaded work, and its Radeon 8060S GPU and 128GB of RAM let you explore AI workloads with passable performance while maximizing game and OS compatibility.

The AI Workstation 300 has no trouble keeping Strix Halo cool under load, but we do wish that Corsair’s thermal system were a bit quieter, especially under light usage. There are four fans of varying types and sizes in this system, and it’s noticeably louder under load than the more compact DGX Spark.

There are two big issues facing the AI Workstation 300. One is that Corsair doesn’t control AMD's software quality. AMD still has plenty of work to do to match the stability and maturity of Nvidia’s CUDA stack for AI workloads, and the performance and reliability of partner systems like the AI Workstation 300 are entirely at the mercy of those efforts in the meantime.

ROCm is improving, to be sure, and AMD’s efforts to boost software compatibility have borne fruit of late. Witness the native AMD ComfyUI build that recently became available for just one example. But that improvement is an ongoing process, and it means you might run into stability and compatibility issues that you won’t have on Nvidia platforms.

The second issue facing the AI Workstation 300 is its price tag. When we first started testing this system, the street price of the config we received was just $2000 or so, which made its performance shortcomings against Nvidia’s DGX Spark much easier to forgive. But that price tag has since spiked to $3000.

That spike is likely thanks to AI-induced NAND and DRAM shortages, and the entire PC market is feeling that pain, not just Corsair. But at that price, the AI Workstation 300 lands quite close to some 1TB SSD configs of Nvidia GB10 systems, like Asus’ Ascent GX10 or Gigabyte’s AI Top Atom. Corsair includes much more storage for the money in our tested config than those GB10 boxes, but we don’t think that should make or break one’s buying decision for a local AI workstation.

You can add a 4TB M.2 2242 SSD like Corsair’s own MP600 Micro or MP700 Micro to an Asus Ascent GX10, for example, and still come out $500 ahead of the $4000 DGX Spark Founders Edition if non-volatile storage space is paramount. Even at $500 more than the AI Workstation 300, we’d consider that a cost worth paying if productive AI development is your primary interest.

GB10 is both a better and more consistent performer for local AI than Strix Halo thanks to both its fundamental architecture and the maturity of Nvidia’s CUDA software stack, and you’ll enjoy that superior performance and stability every time you put a GB10 system to work. If GB10 systems become more expensive due to the same RAM and NAND crunch, however, the AI Workstation 300’s competitive position will benefit accordingly.

All told, if you need a compact PC with a potent x86 CPU, a solid enough GPU for both gaming and some AI exploration, and enough RAM to handle giant AI models, along with flexible storage options to cover Windows and Linux installations alike, the AI Workstation 300 is still a fine platform.

But unless you’re looking for those things at the absolute lowest possible cost and are willing to sacrifice some performance and software maturity in the bargain, the arrival of Nvidia’s GB10 makes Strix Halo—and the AI Workstation 300—tough to recommend at its current price.

Minisforum let me go hands-on with its CES 2026 Innovation Award-winning Mini Workstation, so to get a grasp of what the diminutive new Minisforum MS-02 Ultra could offer, I decided to see if it could replace my aging but capable desktop PC.

This is a true David vs Goliath battle, at least in terms of physical presence. The sub-5-liter MS-02 Ultra battles with my much larger desktop system packed into a 55-liter Fractal Define S PC case. However, the result of the clash wasn’t as clear-cut as I expected, making me wonder what kind of PC I should be using daily in 2026.

Before we get into a specifications table, where you can see exactly how these two systems compare feature-by-feature, it is important to highlight the physical differences. The MS-02 Ultra is an incredible 11 times smaller than my Define S-desktop system. Obviously, such a small size means there are sacrifices to be made in terms of expandability, configurability, thermals, and noise. But those limitations are counterweighted by the benefits of the device being much physically smaller. Dare I say, the Minisforum is kind of portable, and even cute, which is not something I could say about the Define S PC.

The table below makes clear the specs of the Minisforum MS-02 Ultra workstation against those of the system it will have to battle on my desk.

CPU and GPU showdown

The Intel Core Ultra 9 285HX should be much more capable than the i5-14400F that I bought at Amazon for $115 last year. It tempts with the advantages of its modern Intel architecture, higher clock boost, and easily outguns my Raptor Lake Refresh chip on core count. Moreover, the new platform brings quite a lot of modern niceties. Particularly welcome, from my perspective, are the trio of really fast USB4 (including 2x USB4 v2) ports, all with DisplayPort capabilities.

A distinct advantage I expected in moving to the MS-02 Ultra was from the increased core count using Intel’s more modern architecture. Indeed, CPU-heavy benchmarks would show the new chip could convincingly eclipse the old mid-range Core i5-14400F in both single- and multicore workloads, given sufficient power supply and cooling capacity. But, we'll see how that translates to ‘Mark’s world’ of computing in the benchmarks section, below.

I’ll stress again that a brief check of the specs, ahead of any actual comparison testing, may suggest the little Minisforum would win hands-down in CPU-focused processing tasks, but the best consumer-grade GPU I could get to fit inside its sub-5-liter chassis would be crushed by the triple-fan behemoth in my ATX desktop. Indeed, that is sort of what happened. There are nuances, though, especially as I don’t really play any AAA or eSports PC games that launched in the 2020s, or dabble in local AI.

The MS-02 Ultra runs using the integrated graphics on the Core Ultra 9 285HX by. default. However, there’s a PCIe5.0 x16 slot, and incredibly, there’s room to install a two-slot low profile (LP) graphics card. This, plus the spare 8-pin connector from the built-in 350W PSU, allowed me to install a Gigabyte RTX 5060 OC LP GPU.

Nvidia’s RTX 5060 seemed to be the best graphics card choice for me, as a generalist/hobbyist PC user, though I’ve seen other early adopters of the MS-02 Ultra install cards like the Intel Arc Pro B50 (16GB), Nvidia RTX Pro 4000 Blackwell SFF (24GB), and the RTX 4000 SFF Ada (20GB). These offer far more VRAM but can cost a lot more, too. But thanks to this mini workstation allowing for GPU upgrades, perhaps the RTX 6060 or RTX 7060 will arrive one day with more than 8GB of VRAM on board, and in a low-profile form factor. Or maybe AMD will make a performant LP form factor card in the meantime.

More on specs

I chose my Fractal Define S-housed desktop's motherboard as it was one of the few B760 chipset boards with a USB-C 20 Gbps port as standard at the price point. So, getting three times more USB-C ports, with two of them supporting up to 80 Gbps as standard, on the compact MS-02 Ultra was a welcome platform boost. I’m sure there is potential here for some very useful docking or even eGPU use-cases.

I have a pair of Crucial X10 Pro 4TB devices for keeping games, TV, and movie libraries portable. My testing confirmed the USB4 v2 ports on the front work at full USB 3.2 Gen 2×2 (20 Gbps) speeds with the X10 Pro. Some USB4 controllers surprisingly don’t support Gen 2x2 and would fall back to 10 Gbps mode. My X10 Pros run about a third of the speed of the built-in M.2 SSD provided by Minisforum, but I feel they are good for external, highly portable units.

An extra two 25 GbE network ports and two of the available M.2 SSD slots are provided by a pre-installed low-profile card in the MS-02 Ultra’s PCIe 4.0 x16 slot. However, I chose to replace that with a cheap USB Type-A 3.0 card with four ports, as I really like to have lots of USB ports. My old desktop used a USB adaptor for Wi-Fi 6 and didn’t have Bluetooth. The MS-02 has built-in Intel BE200 Wi-Fi 7 and BT 5.4 card.

There are three PCIe slots in the MS-02 Ultra, but adding the GPU takes up the room where two single-slot add-in-cards would otherwise fit. That compares poorly with my Define S desktop, which has five slots. With a bulky GPU fitted in that ATX system, it was still possible to add another three single-slot PCIe cards. Again, I have one installed in there, bristling with extra USB ports.

Preparing for my personal benchmark battle

The most demanding things thrown at my PC are video editing and games like The Division 2 and Far Cry 5. The Cyberpunk 2077 benchmark is just there for an added reference point, (I bought it, but I’ve not actually got around to playing it yet.) The same goes for Shadow of the Tomb Raider. I also like playing in Borderlands GOTY Enhanced Edition, Command & Conquer Generals, and Amiga Forever – but those are so old and lightweight they weren’t worth testing for this performance comparison.

Most of the time, my computer will be used for little more than multiscreen word processing and browsing, communicating with work colleagues, and a touch of ancient Photoshop and Illustrator dabbling. These help me prepare artwork for here on Tom’s Hardware, some YouTube stuff, as well as laser cutting and engraving. Also, Photoshop has recently been replaced by the Affinity suite.

So, without further ado, let’s get into those benchmarks, which unashamedly revolve around my personal general-purpose computer use, with a few gaming classics thrown in. I’ve also included a ‘tweener’ system, where my large desktop was fitted with the incongruously tiny RTX 5060 LP card.

Minisforum MS-02 Ultra productivity performance

The comparison table above doesn’t offer much in the way of surprises. The MS-02 Ultra with Intel’s Core Ultra 9 285HX does exceedingly well in CPU-intensive tests, particularly those that take advantage of the abundance of available threads. We see the 285HX is a star performer in Cinebench R23, with single-thread scores similar to desktop processors like the Core i9-14900K and Ryzen 9 9950X. For multicore, you are also good to compare with the Ryzen 9 7950X3D and i9‑14900F, but the unlocked i9‑14900K starts to pull ahead convincingly.

I also think the 7-Zip (de)compression benchmark shows the MS-02 mini workstation in a good light. Handbrake 4K transcoding wasn’t much better compared with my old 14400F, though. Meanwhile, PassMark Performance Test, a mix of CPU, graphics, memory, and disk tests, showed some worthwhile performance benefits from the Minisforum – when the RTX 5060 was installed.

Minisforum MS-02 Ultra gaming performance

It is almost moot to highlight that the Define S desktop with Radeon RX 9070 would dominate in 3D games and 3D benchmarks. Nevertheless, there were some interesting results showing that the Ultra 9 285HX / RTX 5060 didn’t always beat the i5-14400F / RTX 5060 due to titles simply being ‘GPU-limited.’ An interesting outlier was the crazy-town USA-set Far Cry 5, where Ubisoft’s Dunia Engine 2 (a modified CryEngine) seemed to prefer the desktop Raptor Lake Refresh under the hood. All the games were tested in 1440p, with no upscaling or frame-gen tech applied.

Minisforum MS-02 Ultra power consumption and noise

Now we move into an area of both strength and weakness for the Minisforum MS-02 Ultra. In my benchmarks table, above, I also included power and noise readings. Starting with the good news, the MS-02 requires less power while idling and doing light tasks – like writing this. It also uses about 120W less when gaming, and all the titles I enjoy play smoothly enough on the RTX 5060 system at high to ultra settings at 1440p without the use of any scaling or frame gen tech.

Some bad news for the mini workstation is that it is more audible when idling and under load, in my experience. Initially, I measured both systems from about arm’s length distance. My Define S desktop uses large Noctua fans and sits under the desk, and in the comparison results, you see that even using the tiny RTX 5060 didn’t increase the noticeable system noise – it is hard to be 100% confident about the 1dBA difference observed.

Having the MS-02 Ultra at arm's length from my seated position, but on the desk top, the noise level was quite high at 48 dBA under load, as per the above chart. However, sitting it under the desk, where the big DIY desktop PC was previously located, reduced the max noise level to 41 dBA. But it's a far smaller system, and when positioned in the space vacated by the Define S, the Minisforum front panel moves beyond arm's reach under the desk. It is necessary to bend quite low from the chair to reach the power button and front I/O, which isn't convenient. That's why I decided to keep it on the desk.

I also note that the Minisforum BIOS has various fan modes (quiet, balanced, and performance) and manual tweaking of the trio of system fans. I ran it at ‘Balanced,’ so there’s probably some tuning that could be done for better noise performance.

Bottom line

In some many ways, the Minisforum MS-02 Ultra held its own against my old home-built Goliath.. The pint-sized challenger does everything I need, and some more. There are clear advantages to be had in CPU-heavy workloads. In GPU-heavy tasks and gaming, there’s clearly enough juice for my particular needs today. If you think about the future, then unlike many other mini PCs, this one can get upgraded with a faster next-generation low-profile graphics card (and other compact PCIe cards) if and when they arrive at an affordable price.

I’ll continue to work on configuring the fans, with an eye on processor and memory temperatures, as the louder fans are the one aspect of the MS-02 Ultra that pokes me in the ribs. It may also be repositioned further back on the desk or under it to scale back its audible presence. The fan noise isn’t high-pitched, and I sometimes forget it, but less is more.

We must also talk about the price of the Minisforum MS-02 Ultra mini workstation. As of the this writing, the Intel Core Ultra 9 285HX model with 2x 25 GbE PCIe SFP+ NIC, 32GB DDR5, plus a 1TB M.2 PCIe4x4 SSD, and Windows 11 Pro is priced at $1,583.90 from Amazon.com. You can also go barebone (no RAM, SSD, or OS) for $1,229.90 using a $300 discount coupon on the page, which could be a great option if you have a spare DDR5 SO-DIMM, M.2 SSD, and can source your own OS. As with all PCs at this time, prices are very volatile.

The Gigabyte GeForce RTX 5060 OC LP 8G I bought to go in this system's PCIe ​5.0 x16 slot is currently $359 at Newegg. This was a necessary extra to indulge in the type and quality of gaming fun that I am used to.

Thus, for the Minisforum MS-02 Ultra 32GB/1TB with RTX 5060 LP graphics card installed, as tested, today's total price would be $1,942.90. Things have gotten a little pricier since I began my tests in early January, but that's the way the PC market is going.

We could talk about the comparison between the MS-02 Ultra build/cost and a desktop, with the latter offering much better value and expandability or upgradeability. However, getting something this performant yet compact, with this level of I/O spec and expandability, is going to be impossible with off-the-shelf parts. In this way, and especially for those with a use for the 2x 25 GbE PCIe SFP+ NIC, this mini workstation is a uniquely attractive package.

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 is reportedly preparing to debut its RDNA 4 workstation GPU offerings for desktops, presumably under the Radeon Pro W9000 family. As put forwaard by Hoang Anh Phu, who frequently obtains inside scoops, AMD is considering using the Navi 48 XTW die for its top-end SKUs, paired with 32GB of video memory, likely GDDR6. As always, this leak shouldn't be taken as definitive, but there's likely some truth to it given the proximity of Computex next month, followed by AMD's Advancing AI event in June.

Radeon PRO GPUs are aimed at workstation setups, rivaling Nvidia's (former) Quadro or (now incumbent) RTX PRO offerings for prosumers. These graphics cards bridge the gap between consumers and server domains, for applications like AI, HPC, DCC, CGI, CAD, VR/AR, and the list goes on.

It seems that AMD is sticking to more conservative figures for its flagship workstation offerings this generation. That's somewhat expected since Navi 48 (356mm²) is in the same ballpark as GB203 (378mm²), found in the RTX PRO 4500 Blackwell. Nvidia's top-end GB202 at 750mm², is home to the RTX PRO 6000 Blackwell featuring a massive 96GB frame buffer.

Navi 48, with its 256-bit interface, enables either 16GB of memory (via eight 32-bit channels) or a theoretical maximum of 32GB in clamshell mode, which is the exact configuration being reported here.

Despite severe price hikes to the broader VMWare portfolio, which many companies say are akin to 'being held for ransom,' Broadcom announced today that its two WorkStation Pro and Fusion Pro editions of its VMWare WorkStation Player software would also pivot to a free model. As such, the division between VMWare versions will be solely based on the functionality you want instead of the features you're willing to pay for.

VMWare Fusion Pro and Workstation Pro are no longer available for purchase. Existing commercial contract VMWare users with active support from VMWare have been assured that active support will continue until the end of the contract. However, ticket support through that channel will become impossible afterward.

However, VMWare users needing support won't be out of luck once the premium support disappears. VMWare highlights its existing knowledge base articles, "extensive product documentation," and access to dedicated forums for VMWare Fusion and VMWare Workstation. They also claim that they will still be invested in goals of "new features, usability improvements, and other valuable enhancements" including "maintaining our high standards for stability, with timely updates and reliable performance" while maintaining a focus on "customer-centric growth".

If VMWare is genuinely confident that it doesn't need a premium support tier in the long term, continuing to build upon its existing resources and carefully monitoring its community forums for unfixable bugs and other issues will be integral to its success. Fortunately, the software has been in development for well over 20 years, so most work on it likely boils down to maintenance accounting for new hardware rather than severe bugs or gaps still being ironed out after all this time.

That said, VMWare does have a slightly mixed reputation, with many PC enthusiasts recommending alternative virtualization software. However, the power of VMWare for nested virtualization scenarios is exceptional— late last year, a user of VMWare WorkStation managed to get five Microsoft Operating Systems running simultaneously. This is hardly a standard use case but an impressive demo of VMWare's capabilities now that it's moving into an accessible model for the future.

Hopefully, none of those VMWare Pro commercial contract users are overly reliant on active support for long-term stability in their respective workloads. As long as they're taken care of and left with enough resources to fill the gap once the contract ends, this story seems like a win for most VMWare users, particularly those who want to continue growing the community.

Builders often don't pay enough attention to the plethora of case options available. Or they push the decision off until the end of the component-choosing process. But if you care at all about cooling or how your system looks, I'd advise considering your case options early when planning your build.

The best PC case for you will dictate the system's size, shape, and looks, and what fits inside. The PC case is also a major factor in how quiet your system will be (or how noisy it can get) under load, thanks to any included fans, plus materials like metal mesh or glass panels. These aspects also hugely impact the cooling potential of your rig.

If you want a compact, space-saving PC, consider one of the best Mini-ITX cases to minimize your PC’s footprint on your desk. Or you can go big and get a chassis like Corsair's Obsidian 9000D or Be Quiet's Dark Base Pro 901, to house a massive workstation or gaming rig with acres of expansion possibilities. But really, you don't need a massive case these days to house the best high-end components.

Please check out our Computex 2026 coverage, where we met with several vendors with upcoming chassis, including Cooler Master, NZXT, Hyte, and others.

We've also started fan testing at Tom's and have published our first Best PC Case Fans article covering several recent PC fans including some from Noctua, Arctic, Phanteks, and more. Find out which fans perform best for your needs, be it quiet operation, high-performance, or RGBs, we have you covered and expanding testing testing more fans as we speak.

Below, we’ve selected the best PC cases we've tested, from the dozens of models we've reviewed in recent years. So long as you check to make sure the parts you want to use will fit in it and you like how it looks, one of the cases below should keep you and your PC parts happy for years to come.

Best PC Cases

The Best PC Cases You Can Buy Today

Best-Looking PC Case

We aren't sure who asked for a PC case that mixed mid-century modern design (wooden slats and all) and high airflow, but we sure are glad that Fractal Design answered with the excellent North mid-tower. Available with either a mesh side (as we tested) or tempered glass, and in white with chrome accents on the IO and feet, or black with some seriously pretty brass, this is arguably the prettiest PC case ever produced, period.

And thanks to its pair of 140mm fans and an included side bracket (for the mesh model) that can hold another pair of 140mm fans (not included), the North has excellent thermals as well. Just note that it can get noisy when delivering those temps. The glass side panel may help a bit on the noise front, but likely at the cost of at least some thermal performance.

If you don't like the case's looks, then the company's Meshify 2 Compact (below), is probably a better option for you. But if you're after a case that matches the pretty furniture in your living room, has lots of room for storage, and won't break the bank at $130, nothing else comes close to the North. We hope Fractal's bold move inspires other case makers to try something different.

Read: Fractal Design North Review

Best Budget ATX Gaming Case

Lian Li's Lancool 207 ATX PC case delivers a lot, including some unique features, for just $80 – or $85 if you want the white model. It ships with four fans (two RGB and two uniquely blowing directly on your GPU) and a whole lot of mesh, so airflow shouldn't be a concern. There's also space for large graphics cards and a pair of SATA drives, which should be enough for most of us in this modern M.2 world. Cable routing is also well-designed, making building and cleaning up the interior easier.

The sideways-mounted power supply setup is a unique design choice that mostly works well, but does limit your ability to use large power supplies (160mm in length or less). Aside from that, the only thing I don't like about this case is it's one of the few new models we have tested in 2024 that doesn't support rear-connector motherboards. If that's a feature you need, check out the also-excellent Phanteks XT Pro Ultra.

Read: Hands-on with Lian Li's Lancool 207 PC case: Rethinking ATX on a budget

Best Budget Back-Connect ATX Case

Not long ago, a PC case with four 140mm RGB fans and a tempered glass side panel would have cost well north of $100. But Phanteks manages to deliver that, plus support for high-end components and emerging rear-connector motherboards, in a choice of black or white, for just $80. It's a great case for builders looking to put together a gaming rig with great cooling, who want to spend most of their money on core components like the graphics card and CPU.

The four included RGB fans deliver lots of airflow, and with some tweaking of settings in your BIOS or software, they can also run reasonably quietly. And when I installed an MSI Project Zero motherboard in the case, all the holes lined up perfectly and there was enough room for cable slack. SATA drive support is also good for a mainstream case, with a 3.5-inch drive cage and up to five 2.5-inch drives supported.

It's clear that Phanteks had to cut some corners to keep the costs down (there's a even a stepped-down XP Pro model with a single fan that has an MSRP of just $50). There are no grommets around the cable holes, and only two top USB ports (one USB-C and one USB-A). But for just $80, the Phanteks XT Pro Ultra delivers great airflow, lots of RGB, and support for modern features and high-end hardware. There are better cases out there if you're willing to spend significantly more than $100. But in its price range, Phanteks XT Pro Ultra is hands-down the best option for most gamers and system builders.

Read: Phanteks XT Pro Ultra PC case: Modern budget case delivers lots of airflow and RGB

Best PC Case With a Screen

Lian Li’s Lancool 207 already has a spot on this list for its sub-$100 base model, which features four fans for great thermal performance. But the Lancool 207 Digital adds a six-inch, 1,600 x 720 display up front that is nice and bright, without adding much to the cost of what was already a great case. So at around $105 at this writing, the 207 Digital is easy to recommend as the best PC case with a screen.

Our testing showed the 207 Digital delivers great thermal performance, even though it lacks an exhaust fan. The two fans on the bottom help cool your GPU while gaming, keeping its fans running at a lower RPM, which helps with overall system noise. An included anti-sag GPU bracket makes sure that your graphics card doesn’t put undue stress on your PCIe slot. We also found the screen to be bright, crisp, and versatile since it can display system stats via Lian Li’s software, or function as a secondary screen in your OS of choice, displaying literally whatever you like. If you’re after a case with a screen – especially if you still prioritize cooling performance and don’t want to pay more than typical mid-range case prices, the Lancool 207 Digital should be on your (very) short list.

Read: Lian Li Lancool 207 Digital review

Best Mainstream Airflow-Focused Case

With just two of the company's F120Q Airflow case fans behind a skeletal frame with a perforated front panel, the NZXT H7 Flow manages to deliver excellent temperature and noise level numbers in our testing, beating or competing well against pricier chassis with more fans and glass panels. It's also a pleasure to build in, with good cable management and a $130 MSRP that's quite reasonable in this era of seemingly ever-increasing prices.

The primary missing feature here, which will feel refreshing for many builders, is the lack of any RGB to light up your system. That, combined with the boxy look and perforated front means this case isn't the prettiest on our list, but it can certainly look good with some tasteful lighting delivered via your components, extra fans or a light strip. And if you don't like the look of the fully vented front and you aren't that concerned about keeping your temps a low as they can go, the non-Flow H7 model has a solid front, but with everything else we loved about the H7 Flow.

Read: NZXT H7 Flow review

Best Performance-Focused ATX Case

Lian Li's excellent Lancool 217 borrows from its Lancool 216 and Lancool 207, while adding some tasteful wood trim. It also sports five fans, with two massive 170 mm spinners up front and dedicated GPU intake fans above the PSU chamber to help chill today's best graphics cards. With all that airflow and a mesh front, the Lancool 217 is tough to beat in its $125 price range.

The 217 was also a pleasure to build in, thanks to a fused front-panel connector block, ample cable routing, and a GPU support bracket. Back-connector motherboards are also supported for an ultra-clean build, and depending on how you configure it, there's support for up to five SATA drives.

The only real negative about this case for me is that its front-panel ports are on the bottom of the case (on the side near the front), which doesn't seem convenient for a case that's likely to live on the floor (it's roughly 20 inches tall and deep).

Read: Lian Li Lancool 217 hands-on

Best Colorful PC Case

Hyte’s X50 is in an aesthetic realm all its own, with its rounded design and several striking colors: purple, pink (as tested), green, red, black, or white. The build quality of the X50 is second to none, with higher-quality construction than its competitors. At $159.99 in the United States, it’s also not super expensive, although keep in mind the case ships without any fans. If you want to add color-matched fans, Hyte sells a four-pack for $49.99, or $10 less for white or black fans. And there is no RGB here (although Hyte also sells magnetic light strips), should you want to add those as well.

While it isn’t cheap (and indeed it feels premium), for those looking for a softer case design, or just something more colorful than most builds, the X50 is easy to recommend. It’s not the best performer, but our testing didn’t kick up any red flags when it comes to cooling. If temps are a concern, you can add more than the four fans we used for testing (the case supports 10), which would likely help keep things running even cooler.

Read: Hyte X50 review

Best Case for Rear-Connector Motherboards

Corsair's 2500D airflow doesn't stand out much as a general Micro ATX case. But as a case specifically for rear-connector motherboards, it's the best I've tested so far. And if you're after a case for a full-size ATX build, the company also sells a 6500D Airflow for $199. These cases get the basics right, like cutout placement (although a PSU support strip right above the bottom cutout is annoying). But more importantly, due to their dual-chamber design, they provide enough room for the large amount of cable slack that results from plugging nearly everything in behind the motherboard, instead of up front. The other cases I've tested with these rear-connector boards don't have nearly enough space behind the motherboard. For even more room for cables on the 2500D, you can opt for a smaller SFX power supply. But of course, that generally costs extra.

I do find the basic presentation of the 2500D a bit drab, but adding the $69 wood accent panels morphs the 2500D from a plain bulky box to something I enjoy looing at. That's not going to be worth the extra expense for everyone, but because the panels are accessories, you don't have to decide whether you want them or not until after you own the case. And you are spoiled for choice, as the company offers panels in silver aluminum and four wood varieties. Other accessories include a vertical GPU mounting kit and a front panel to turn the 2500D into the 2500X.

Read: Hands-on with Corsair's 2500D Airflow case

Best Design-Focused Mini-ITX Case

The substantial size and substantial cooling needs of today's components – particularly graphics cards like Nvidia's RTX 4090 – have made several compact cases from the past few years effectively obsolete. But at just 10.4 liters of volume, the Mini-ITX Fractal Design Terra case has enough room for most big GPUs, while delivering attractively thanks to an aluminum exterior with some attractive wood on the front.

It doesn't have as wide of an appeal as its ATX North cousin, since you'll be far more limited in your component and airflow options with the Terra. But if you don't need to pack in the absolute most powerful CPU (cooling clearance is particularly cramped) and you don't need several storage drives, the Terra is compact, practical, and stunning – especially if you want a respite from PC gaming's typical RGB and aggressive designs. The Terra's flip-up doors and adjustable spine also make this case surprisingly easy to build considering it's so compact. Just know you'll need to bring a modular SFX PSU, because you'll need all the spare interanal space you can get with this case.

Read: Fractal Design Terra Hands On

Best Premium

The Cooler Master Cosmos Alpha stands out as a premium full-tower case, thanks to the refined, modern design, massive internal space, and the modularity of its design. You can move fan mounts and use fans of different sizes, and slide the motherboard tray to adjust layouts to fit high-end hardware. Supporting up to four radiators, 12 fans, and a wide variety of motherboard sizes, the Alpha has loads of cooling potential. It also has plenty of room for cable management and modern front-panel connectivity with multiple USB Type-A and Type-C ports (3.2 Gen 2x2 20 Gbps). Premium touches such as the aluminum frame, signature handles, and a hinged smoked-glass panel round out its flagship aesthetic.

The case is heavy and requires a lot of space, plus the 2.5/3.5-inch drive capacity feels limited for the size. But these are minor complaints for a case that excels in flexibility, cooling, and long-term usability. The combination of thoughtful design, build quality, and extensive hardware support makes the Cosmos Alpha the best choice for enthusiasts seeking a high-end, long-lasting, premium tower for large motherboards and powerful hardware.

Read: Cooler Master Cosmos Alpha Hands-on

Best Large "Fish Tank" PC case

NZXT’s 2025 H9 Flow has a sleek design, without RGB bling. It's spacious inside, with lots of room to work with on both sides of the case. The main chamber supports the largest GPUs on the market and dual 420 mm radiators, while the back has extra space for cable storage.

Liquid cooling enthusiasts will be pleased to know the H9 Flow supports two 420mm radiators, along with its angled intake fans, for the best possible temperatures. And storage fans will appreciate the support for up to six SATA drives – the most we’ve seen on a mainstream case in quite a while. The H9 Flow brings a lot to the table in terms of features and support, and it doesn't lean too heavily on your bank account, selling for $169 when we wrote this.

Read: NZXT H9 Flow (2025)

Best Case for High-End Rigs and Workstations

Fractal's Meshify 2 is well-thought-out and versatile, delivering an enjoyable building experience. Whether you use this case as a system where you just want to deliver copious amounts of airflow and expansion space, a workstation with a cadre of hard drives, a server, or high-end custom liquid cooling, the Meshify 2 can accommodate your build. Its thermal and acoustic performance is also in line with what we'd expect from a mesh front. All this cobined earns this case a rare five-star rating.

The only thing to consider before hitting the buy button is whether you really need the space for storage or cooling parts, or should you buy a smaller chassis and save some money and sapce?

Read: Fractal Design Meshify 2 Review

Also tested

Not every case we test earns a spot on our best list. Whether it be for a lack of features, high price, poor performance, or something else, the cases below aren't the best we've tested. But the best PC case for you has to be something you like to look at. So the cases we've tested below may be worth considering, especially if you find them on sale.

Cougar Airface Eco:

Cougar’s oddly named Airface Eco case features a decent all-around design, support for large GPUs in vertical orientation, a modular IO panel, and a built-in GPU holder. It also sports a distinctive front face with large holes, but doesn't otherwise stand out in terms of features or price. You can read more in our full Cougar Airface Eco review.

Be Quiet Dark Base 701:

Be Quiet’s Dark Base 701 is a high-quality. customizable case that offers one-touch fan and lighting controls from the top panel, and it performed well in our tests. It supports vertical GPU mounting, ample SATA storage, and custom liquid cooling reservoirs.

Read: Be Quiet Dark Base 701 review

Montech Air 903 Max:

This case doesn't do anything new when it comes to design or features. But with large motherboard support, good airflow thanks to four include fans, and a $75 price, there isn’t a whole lot to complain about.

Read: Montech Air 903 Max review

APNX V1:

This case combines angled front fans and optional wood accents, delivering good thermal performance and a reasonable price. But its looks aren't for everyone, and its screwed-down panels feel a little dated.

Read: APNX review

Be Quiet Light Base 900 DX: This case's solid build quality and attractive aesthetics are paired with the somewhat unique feature of having three different layout options. Thermal performance is middle of the road, which is to say while it might not be the best, it will be good enough for 97% of users looking to build a PC for gaming or mixed use. While there are multiple fan hubs, you'll have to supply you own fans.

Read: Be Quiet Light Base 900 DX review

MSI MEG Prospect 700R:

At aropund $380, the MEG Prospect 700R from MSI delivers a lot of features, like a top-mounted touchscreen for supporting various functions. Hinged doors make getting inside the case easy, and it is one of a limited number of chassis that allow fancy PSUs to show their displays.

But the case would be much easier to recommend if it didn't cost so much. In particular, it should include a GPU support bracket for horizontal mounts and a PCIe riser cable for vertical mounting, considering its nearly $400 asking price.

Quick PC Case Shopping Tips

MORE: Best PC Cases

MORE: Best Mini-ITX Cases

If you're looking for the best CPUs for gaming or one of the best budget CPUs, there are only two choices: AMD and Intel. That fact has spawned an almost religious following for both camps, and the resulting AMD vs Intel flamewars make it tricky to get unbiased advice about the best choice for your next processor. But in many cases, the answer is actually very clear: AMD's X3D chips dominate gaming performance, from the Ryzen 5 7600X3D up to the Ryzen 9 9950X3D2. For application performance, AMD still tops the charts with flagships, but Intel's recent Arrow Lake Refresh chips provide nearly the same level of performance under $300 based on the data in our CPU benchmark hierarchy.

This article covers the never-ending argument of AMD vs Intel desktop CPUs (we're not covering laptop or server chips). We judge the chips on nine criteria based on what you plan to do with your PC: pricing, specifications and features, gaming performance, productivity performance, driver support, power consumption, overclocking, chipset and socket support, and security, giving us a clear view of the state of the competition. Throughout each section, we'll also discuss the process nodes and architectures influencing the moving goalposts. However, each brand has its strengths and weaknesses, so which CPU brand you should buy depends mostly on what blend of features, price, and performance is important to you.

The latest 'Arrow Lake Core Ultra 200S series builds on Intel's hybrid architecture formula but leverages a new design philosophy that constrains gaming performance. However, the chips do provide strong performance in single- and multi-thread productivity workloads, along with improved energy efficiency. New refreshed chips like the Core Ultra 7 270K Plus shore up the gaming front and provide a big boost in productivity performance, as well.

AMD's answer is its Zen 5 Ryzen 9000 series. These chips debuted with a fizzle on the gaming front, but subsequent firmware and operating system updates have improved the overall performance significantly, making these chips the bang-for-the-buck leader for most users looking for a balanced system.

AMD's powerful gaming-optimized Ryzen 9000 "X3D" models utilize a breakthrough 3D chip-stacking tech to take the lead in gaming performance, as you can see in our Ryzen 9 9850X3D review that finds the 9850X3D to be the fastest gaming CPU on the planet, bar none. AMD has also leveraged this X3D tech with its older architectures to make incredibly value-centric gaming chips, like the Ryzen 5 5600X3D. The gains in gaming performance are phenomenal, so much so that the competition against Intel chips isn't even close, giving AMD a walk-in touchdown for gaming PCs.

AMD vs Intel: Which CPU is Best?

Comparing Intel vs AMD in 9 categories in 2026, AMD and Intel are in lockstep with a win in three categories each, with the brands coming in tied in two categories. Despite long-standing as the value alternative, AMD’s original Zen microarchitecture has kicked off a slow rise to the top. The Steam hardware survey now shows AMD only slightly behind Intel in market share, and from the Passmark database, AMD outpaced Intel in desktop market share for the first time last year.

Although the earlier years of Zen were met with an audacious Intel, that story quickly changed as AMD started to gain ground, particularly around the release of Zen 3. Intel was finally able to move onto its 10nm node after years of manufacturing challenges, as well as introduce a hybrid architecture in a desktop x86 CPU. Alder Lake, Raptor Lake, and Raptor Lake Refresh all maintain competitive positions against their Zen 3, Zen 4, and even Zen 5 counterparts, but Intel’s market dominance – and the price it demanded from it – has disappeared in desktops.

Fierce competition is good for everyone, but Intel didn’t maintain its trajectory, instead opting for a radically different architecture with Arrow Lake that ditched Hyper-Threading. In AMD’s camp, it doubled-down on the wildly successful 3D V-Cache, introducing a second generation of the technology that places additional cache below the compute die for more thermal headroom, and as a result, higher clock speeds.

Although Arrow Lake brings much-needed efficiency improvements to Intel’s desktop line-up, as well as great overclocking headroom on CPU cores and the memory subsystem, AMD dominates gaming performance with its X3D lineup, sometimes by as much as 30%. At the same time, AMD is able to match or exceed Intel’s productivity performance, despite how impressive the Arrow Lake architecture is from a technical perspective.

Broadly, AMD earns our recommendation for most people. It clearly leads in gaming performance with X3D CPUs, and despite stagnant specifications, AMD has continued to show architectural advantage in comparative performance scenarios. Intel has gained by some ground with the 270K Plus and Core Ultra 5 250K Plus, which are both excellent processors are affordable prices. For now, however, AMD remains our go-to recommendation.

AMD vs Intel CPU Pricing and Value 2025

AMD has long stood as the value king, and although you can get a lot for your money with chips like the Ryzen 7 9800X3D, Intel has slid into the value position as AMD continues to rise in market share. Looking at like-for-like CPUs in the AMD vs Intel battle, the two are in lockstep on pricing within a margin of $20 to $30 in most cases, outside of sales.

For the current generation lineup, Intel’s Arrow Lake chips are slightly cheaper at the Core Ultra 9 and Core Ultra 7 levels, though only by $20. Factoring in sales, Arrow Lake ends up much cheaper. At the time of writing, for instance, the Core Ultra 9 285K is available for $500 while the Ryzen 9 9950X will run you $570.

The prices above are based on the average selling price of the main lineups from AMD and Intel over the past 60 days. Notably, these aren’t list prices, which aren’t reflective of what you can expect to spend, especially this late into a generation. As an example, the Core Ultra 7 265K launched at $395 while the Ryzen 7 9700X launched for $359. Now, however, Intel comes out slightly cheaper.

AMD leads in overall value for lower-end offerings like the Ryzen 5 9600X, but Intel isn't far behind with the recent Core Ultra 5 250K Plus. In gaming performance, you can see Intel is delivering better average frame rates, as well, though only with its newer Arrow Lake Refresh chips.

AMD and Intel flip positions when looking at last-gen options, with only the Core i9-14900K coming in slightly less than the Ryzen 9 7950X. Intel has clearly recognized the value proposition of the Core i5-14600K and Core i7-14700K, with both now selling for around the same price as their Arrow Lake counterparts.

Winner: Intel. Intel and AMD are more competitive on pricing than they’ve been in years. Intel comes out slightly ahead on value with lower-end SKUs, but value demands context. The Ryzen 7 9800X3D might be a worse value than the Core i5-14600K, but if you have $480 to spend on a CPU and want the best gaming performance, AMD’s X3D chip is right for you.

However, Intel shook things up massively with the 270K Plus and 250K Plus. Now, Intel is offering competitive gaming performance around $200 to $300 while pushing flagship-level application performance. For now, at least, Intel is in the driver's seat when it comes to overall value through Arrow Lake Refresh.

AMD vs Intel CPU Gaming Performance 2025

In the AMD vs Intel CPU gaming performance battle, AMD's X3D processors hold the lead in all critical price bands. Below, we have a wide selection of collective gaming performance measurements for the existing chips in the different price bands. You can see a much more holistic view in our CPU Benchmarks Hierarchy.

As you can see, AMD's standard 'Granite Ridge' Ryzen 9000 models take the lead over Intel's current-gen Core Ultra 200S 'Arrow Lake' processors, and the gaming optimized Ryzen 9000X3D chips, which cost more than the standard models, take an overwhelming lead, often by 30% or more.

Intel's current-gen Core Ultra 9 285K is the new flagship, but its unique tile-based (chiplet) design has a negative impact on gaming performance, so it isn't as fast as the prior-gen models in gaming. That leaves the Core i9-14900K as the fastest Intel gaming chip, and it often sells at a steep discount. The recent Core Ultra 7 270K Plus, however, is just marginally behind the Core i9-14900K, and it's available for around $350.

AMD's standard Ryzen 9000 models, like the flagship Ryzen 9 9950X, are faster in gaming than Intel's Core Ultra 200S series, but Intel's previous-gen 14900K holds the edge. However, AMD's X3D models, like the Ryzen 7 9800X3D and the Ryzen 9 9950X3D, are the fastest gaming chips on the planet, beating Intel's current-gen chips by 30+%. These chips also deliver the full performance of the standard models in productivity workloads, eliminating the tradeoffs associated with the X3D models that we've seen in the past. The X3D models do come at a premium, but they justify their price tag with the fastest gaming performance money can buy.

Older X3D chips leverage AMD’s first-gen 3D V-Cache, which stacks the cache on top of the compute die rather than below, which is the case with the latest Zen 5 X3D offerings. That design limits thermal headroom, and in turn, clock speeds, so expect a dip in productivity performance if you aren’t using one of AMD’s latest X3D offerings.

Kicking your resolution up to 1440p and beyond typically pushes the bottleneck back to the GPU, so you won't gain as much from your CPU's gaming prowess. However, a bit of extra CPU gaming performance could pay off if you plan on updating your graphics card (see our best graphics cards) with a newer generation while keeping the rest of your system intact. We expect most builds in the mid-range to come with lesser GPUs, which generally serve as an equalizer in terms of CPU performance.

In terms of integrated graphics performance, there's no beating AMD. The company's current-gen Phoenix Point APUs offer the best performance with the 8700G and 8600G. We’ve seen even more capable APUs out of AMD with its mobile Strix Point offerings like the Ryzen AI Max 395, but Strix Point hasn’t made its way to desktop yet.

Winner: AMD. AMD's standard Ryzen 9000 models deliver solid performance in gaming, beating Intel's current-gen models, but they trail the previous-gen Intel chips. However, the Ryzen 9000X3D models take the unequivocal lead in gaming by massive margins, giving AMD an easy win in the gaming department. The previous-gen Raptor Lake Refresh models are a distant second in gaming performance, and they aren't as performant as the Ryzen 9000 models in productivity applications, making them a tough choice as an alternative.

AMD vs Intel Productivity and Content Creation Performance 2025

In the non-gaming performance battle of AMD vs Intel CPUs, Intel's Arrow Lake chips have made great strides against AMD's finest, but they fail to deliver a tangible enough lead over AMD's core-heavy models to make productivity performance a make-or-break buying decision. AMD leads in overall multi-threaded performance with its Ryzen 9 9950X3D and 9950X by a slim margin, but Intel keeps a solid grip on the lead in single-threaded applications. The new Arrow Lake Refresh CPUs don't radically change the performance picture overall, but they come in much less than AMD's Zen 5 competition.

Arrow Lake marks Intel's continuation of using a mixture of two types of cores for the mainstream desktop PC. The big performance cores (P-cores) are best for latency-sensitive work, giving Intel the uncontested lead in single-threaded applications. The efficiency cores (E-cores) step in to add some additional heft in threaded and background applications, which pays big dividends in heavily-threaded content creation and productivity applications.

In contrast, AMD continues to leverage its tried-and-true P-core only design, which also supports AVX-512, a big consideration for those who leverage powerful productivity applications that employ the high-performance instructions. Zen 5 brought a 512-bit data path for AVX instructions – Zen 4 used two 256-bit paths – which provides a massive performance advantage in applications that leverage AVX-512, even compared to Zen 4. Intel’s latest CPUs don’t support AVX-512 instructions.

Solid performance in single-threaded work equates to faster performance in all manner of workloads, particularly day-to-day applications that rely on snappy responsiveness from the processor. The Core 7 270K Plus has taken the uncontested lead in single-threaded performance across the full suite of our benchmarks, and it comes in at a mid-range price point. And on the multithreaded front, Intel is able to match the Ryzen 9 9950X despite the Core Ultra 7 270K Plus coming in at half the price.

Winner: Intel. Although AMD offers compelling productivity performance with its most expensive offerings, the Core Ultra 7 270K Plus and Core Ultra 5 250K Plus punch far above their weight class in productivity performance, easily delivering flagship performance at entry-level prices.

AMD vs Intel Processor Specifications and Features 2025

Modern Intel and AMD CPUs sport much different architectures, so comparing specifications directly is a fool’s errand in most cases. AMD has a traditional approach with a homogeneous core architecture that sports simultaneous multi-threading. Intel moved onto a heterogeneous architecture with Alder Lake, which it has maintained since, and it removed Hyper-Threading with Arrow Lake.

This departure in architecture means you can draw inaccurate conclusions from specs alone. For instance, Intel’s flagship Core Ultra 9 285K only has 24 threads compared to AMD’s 32 on the Ryzen 9 9950X. Despite that, the two chips offer competitive multi-threaded performance, as you can see in our performance results in the above section.

The cores and clocks aren’t as important as they once were, at least when comparing a Ryzen 9 to a Core Ultra 9, or a Ryzen 5 to a Core Ultra 5. Instead, there’s been a stark shift in focus toward cache in the face of 3D V-Cache. That’s the marquee feature AMD has been running on for the past few years, offering chart-dominating performance in games that Intel has yet to answer. Previously, 3D V-Cache came with a slight compromise to productivity performance due to clock speeds on eight-core parts and CCX-to-CCX latency on 12- and 16-core parts. AMD has rectified that issue with second-gen 3D V-Cache, however.

Apples-to-apples, Intel and AMD offer similar cache amounts when excluding X3D parts, though Intel more evenly distributes the cache between L2 and L3, while AMD focuses the lion’s share of cache in the last level. Intel is rumored to pack bLLC into its upcoming Nova Lake generation to answer 3D V-Cache. Intel currently deploys this style of cache packaging in its Clearwater Forest server chips, but it has yet to show up in a consumer CPU.

Winner: AMD. Intel has highlighted several features over the past several generations, from Thread Director to Intel APO and even the new iBOT feature, but these features have mainly served to bring Intel’s unique architectural design up to par. Meanwhile, AMD has pushed ahead with a big win in gaming due to 3D V-Cache, which only improved in its second generation. Add on top of that AVX-512 support, as well as a true 512-bit data path in Zen 5, and Team Red easily wins this category.

AMD vs Intel CPU Power Consumption and Heat 2025

When comparing AMD vs Intel CPU power and heat, the former's 4nm TSMC process node makes a big difference. Power consumption comes as a byproduct of design choices, like lithography and architecture. However, higher power consumption often correlates to more heat generation, so you'll need beefier coolers to offset the heat output of greedier chips.

Overall, Intel has reduced its power consumption from meme-worthy to an acceptable level, but it still consumes more power than Ryzen. Still, in aggregate, AMD's 4nm chips either consume less power or provide much better power-to-performance efficiency.

As you can see above, you'll get more work done per watt of energy consumed, which is a win-win, and AMD's cooling requirements aren't nearly as overbearing. Arrow Lake brings Intel more in-line with AMD, both when looking at peak power consumption and productivity efficiency. Even then, however, Intel’s latest offerings end up slightly ahead of AMD in overall power consumption, and slightly behind in over efficiency.

Winner: AMD. In judging AMD vs Intel CPU performance per watt, it's impossible to overstate the importance of having the densest process node paired with an efficient microarchitecture, and TSMC's 4nm and AMD's Zen 5 are the winning combination. The latest Ryzen processors consume less power than Intel on a power-vs-performance basis, even in the face of Arrow Lake.

Intel vs AMD CPU Overclocking 2025

There's no debate when you compare Intel vs AMD CPU overclocking. Intel offers the most overclocking headroom, meaning you can gain more performance over the baseline speed with Intel chips than you can with AMD's Ryzen processors. To learn more about overclocking, head to our How to Overclock a CPU guide.

Dynamic boosting algorithms and increasingly multi-threaded applications have offset some of the previous benefits of overclocking, at least for core clock speeds. With AMD’s current CPUs, and up to Raptor Lake Refresh for Intel, a modest all-core overclock generally won’t provide a performance benefit, and it could even reduce your performance, though only in lightly-threaded applications. The boosting algorithms can usually achieve higher speeds for these applications, with all-core overclocks running into a thermal wall before providing much of a boost.

However, for Intel, that changed with Arrow Lake. It ships with much higher thermal headroom than the last several Intel generations, shortening the gap between single-core boosts and all-core speeds. Intel now also provides more fine-grain control over clock speeds, with increments as low as 16MHz. Arrow Lake scales particularly well with high DDR5 speeds, offering a small performance edge if you’re able to secure a fast kit of memory.

AMD has taken a different approach. Rather than offering more granular control over overclocking, it has continued to invest in Precision Boost Overdrive (PBO) as a one-click overclocking solution. In addition, AMD offers Curve Optimizer and Shaper, which allow you to easily undervolt your chip with a voltage offset across the frequency range.

Intel has an edge in overclocking, but it comes with a caveat. In order to overclock an Intel chip, you need to purchase an unlocked K-series processor as well as a Z-series motherboard. H- and B-series motherboards only support memory overclocking. Meanwhile, AMD supports overclocking on all of its modern CPUs, short of the Ryzen 7 5800X3D and other Zen 3 chips sporting 3D V-Cache. AMD also supports CPU overclocking on B- and X-series chipsets, saving you some money.

Winner: Intel. Although CPU overclocking isn’t as prominent as it used to be, Team Blue holds the edge in this AMD vs Intel battle. You have more granular control over core overclocking, as well as additional fabric and memory overclocking features that can net higher performance on Arrow Lake. AMD has focused on making overclocking more accessible, but that also leaves enthusiasts with less room to play.

AMD vs Intel CPU Drivers, Software, and Firmware 2025

Over the past couple of years, both AMD and Intel have seen catastrophic firmware issues, which caused CPUs to fail, or in the worst cases, literally burn themselves up. In the most recent generation, we also saw hamstrung performance due to software issues from both AMD and Intel, though both have also released a string of microcode and driver updates to address these problems.

On the driver front, Intel and AMD are both stable. In the early days of Zen, we saw lopsided support for Intel’s latest architecture, and AMD’s smaller position in the market meant issues were more common and fixes took longer to arrive. That isn’t the case today. Broadly, Intel and AMD’s chipset drivers are stable, and issues that do arise are usually resolved through an update within a few days.

Firmware is a different story, and that’s been an unfortunate area of focus for AMD and Intel over the last couple of years. Intel was at the center of a years-long issue with instability on Raptor Lake and Raptor Lake Refresh chips, and in particular, the Core i9-13900K and Core i9-14900K. Reports of instability in games on these CPUs date back at least as far as early 2023, but Intel didn’t acknowledge them until mid-2024.

Intel released a microcode update in late 2024 that solved most instability problems, but as recently as September 2025, it’s continued to work on the issue.

In AMD’s camp, high-end Ryzen CPUs have burned themselves up in the socket, ruining both the chip and motherboard – twice, though for different reasons. In the first go, the Ryzen 7 7800X3D was operating at elevated voltages. An AGESA update shortly after reports started circulating capped the voltage at 1.3V, and there haven’t been any reports since.

With the release of the Ryzen 7 9800X3D, a far larger number of reports started circulating, once again showing the burning on the CPU’s contact pads and in the motherboard socket. This issue was much messier than the original Ryzen 7 7800X3D, though the vast majority of reports came from the Ryzen 7 9800X3D paired with an ASRock motherboard. In this case, AMD shifted blame to ASRock. ASRock has reportedly solved the problem through BIOS updates, but new reports are still trickling out at a reduced cadence.

Winner: Tie. It’s hard to declare a winner in this category given the two major controversies Intel and AMD have faced over the past few years, both of which resulted in CPUs being ruined while supposedly operating within warrantied specifications.

Although it’s important to highlight these problems and the responses to them from AMD and Intel, they aren’t widespread issues. Neither triggered a proper recall.

AMD vs Intel Chipset and Socket Support 2025

An increasingly important aspect of choosing between Intel and AMD is the longevity of the chipset and socket of your motherboard. Although new generations bring new chipsets, AMD has set a new standard for socket support with AM4, which it appears to be carrying forward with AM5. Intel has responded in kind, supporting 12th-, 13th-, and 14th-gen chips on the same LGA1700 socket. However, it moved to the LGA1851 socket with the release of Arrow Lake, and it will move to the LGA1954 socket with the release of Nova Lake CPUs in late 2026.

AMD’s current commitment is to support the AM5 socket through 2027, but that timeframe could extend. With the release of Zen 4 and introduction of AM5, AMD originally only committed to support through 2025 before extending the timeframe with the release of Zen 5. With AMD’s current release cadence, that means we should see at least three generations, and possibly four, on the AM5 socket. If AMD once again extends the window, we’re likely looking at five generations of support.

It’s certainly possible that AMD will extend the timeframe. The AM4 socket set a precedent of longevity. It was originally introduced in 2016, and AMD has released new AM4 chips as recently as 2025; though, these chips are mostly rebranded versions of old silicon, or specific variants for different international markets.

For chipsets, AMD currently supports Zen 4 and Zen 5 CPUs across all chipsets with the AM5 socket. With AM4, AMD eventually brought support for Zen, Zen+, Zen 2, and Zen 3, though support varies based on manufacturer for older chipsets.

Intel has a more confusing lineup of chipsets and support, which alone spoils the reasoning for AMD winning this section. After three generations on the same socket, Intel moved to a new socket with the release of Arrow Lake. That socket (LGA1851) has seen Arrow Lake Refresh chips, and we don't expect it to see any more releases.

Intel also forces customers to its flagship Z-series chipset for CPU overclocking support; a line in the sand that it’s continued to maintain despite the fact that AMD supports CPU and memory overclocking on both B-series and X-series chipsets. Thankfully, the last several generations of Intel chipsets have opened up memory overclocking on B-series chipsets.

Winner: AMD. The AM4 socket brought about a new standard for drop-in CPU replacements, which AMD is bringing forward with AM5. Intel has made strides to support multiple generations on the same socket, but it has yet to demonstrate the ability to deliver support over several years like AMD has. Additionally, Intel continues to charge a premium for CPU overclocking by restricting the feature to flagship Z-series chipsets.

AMD vs Intel CPU Security 2025

The last few years have found security researchers poking and prodding at the speculative execution engine that's one of the key performance-boosting features behind all modern chips. The resulting research has spawned an almost never-ending onslaught of new vulnerabilities that threaten the safety of your system and private data. Unfortunately, these types of vulnerabilities are incredibly dangerous because they are undetectable—these tactics steal data by using the processor exactly as it was designed; thus, they are undetectable by any known anti-virus program.

The rash of fixes required to plug these holes also continues to grow, and many of them resulted in reduced performance initially. AMD and Intel have recognized the cost of such updates during the Spectre and Meltdown era, and more recent patches come with little to no impact on performance. It’s possible that could change with a future vulnerability, but we haven’t seen chart-breaking changes in performance from security updates in the past couple of generations.

Intel still has more known vulnerabilities, but in the years following Spectre, it’s become clear that both AMD and Intel are vulnerable to this family of attacks. At the time of writing, Intel recently patched over 30 security bugs, and AMD confirmed a critical security vulnerability in its random number generator on Zen 5 CPUs, which it plans to address with an AGESA update.

Winner: Tie. Similar to firmware, it’s hard to say anyone is a winner here. Intel claims it surfaces and addresses vulnerabilities more proactively than AMD does, but both brands chase the constantly evolving world of threats and offer driver and/or firmware updates in response. In addition, both maintain a bug bounty program to incentivize security researchers to find and surface exploits.

FAQ: Frequently Asked Questions

• MORE: Best CPUs for Gaming
• MORE: CPU Benchmark Hierarchy
• MORE: Zen 4 Ryzen 7000 All We Know
• MORE: How to Overclock a CPU
• MORE: How to check CPU Temperature
• MORE: All CPUs Content

Alongside pre-launch retailer listings, VideoCardz has unearthed a low-res render of a purported RTX 2000 Ada desktop workstation GPU. The source reckons the release date for this entry-level Ada workstation card with 8GB of VRAM is sometime during the coming week. 

This new RTX 2000 workstation GPU seems to be yet another AD104-based GPU, with 8GB of (likely ECC) GDDR6 VRAM and PCIe Gen 4 x16 support. It also appears to lack any external power connector, so we should expect a TDP of no greater than 75 Watts (the usual maximum power draw of a PCIe slot).

Two model names listed in the VideoCardz report are the Nvidia RTX 2000 ADA and the Nvidia RTX2000E ADA, with estimated price points of $600-650 and $500-630, respectively. What the "E" suffix signified is currently unconfirmed, but "E" may refer to "Enterprise". It is the cheaper of the two models, though, so perhaps that model is only available when purchasing in bulk.

Since the AD104 is the GPU used in the consumer GeForce RTX 4070 and RTX 4080, this workstation card should in theory be quite powerful for its (potentially SFF) form factor and low wattage. Reducing wattage to this extent will greatly reduce performance, but AD104 provides a strong foundation. Additionally, all of Nvidia's workstation (previously dubbed Quadro) GPUs are sold with blower-style coolers, making them more favorable for multi-GPU stacking in servers and other workstation scenarios.

At the time of writing, an entry-level '2000' model is conspicuous by its absence from Nvidia's listings of Ada architecture RTX GPUs for workstations. The purported Ada Generation RTX 2000 GPU specs line up with a long-due refresh of Nvidia's entry-level workstation video card offerings.

With any luck, this rounding out of the Ada Lovelace workstation offerings may also push Nvidia toward considering new branding for its next generation of workstation GPUs. The naming scheme of workstation Nvidia GPUs has gotten increasingly contrived gen-over-gen since the change, particularly since mainstream RTX GPUs use a nearly identical naming scheme, except focusing on a fixed RTX XX-Y0 range instead of RTX X-000 for each model number.

Needing to add 'Ada Generation' to every product listing made for a current-generation Nvidia workstation GPU seems a little excessive for a system that used to use Quadro branding and some numbers more effectively. Still, we'd be veering off topic if we lingered on this any longer. It is worth mentioning now that Nvidia's Ada Lovelace lineup seems complete, though.

Along with its massive core count and extended memory support, AMD's Ryzen Threadripper Pro 7995WX has a rarely-used feature that makes it stand out in the workstation processor market: overclockability. Not many people buy a $10,000 CPU to overclock it, so AMD decided show how capable it is itself. Last week the company's lead overclocker Sampson managed to overclock the CPU to 4800 MHz and 5200 MHz using air and liquid cooling. But this week the CPU was able to hit 6.0 GHz with all 96 cores. 

Experiments in overclocking AMD's Ryzen Threadripper Pro 7995WX processor using custom air and liquid cooling were conducted by Bill 'Sampson' Alverson. Extreme overclocking experiments using liquid nitrogen (LN2) were done by two teams — one composed of Bill 'Sampson' Alverson and Amit Mehra from AMD, and the other featuring renowned overclockers Roman 'der8auer' Hartung and Jon 'Elmor' Sandström. 

The results are quite remarkable. Elmor and Der8auer experiments saw the CPU reach a whopping 6.0 GHz clock speed using liquid nitrogen. Sampson and Amit Mehra increased the frequency of all 96 cores in the AMD Ryzen Threadripper Pro 7995WX to an impressive 5,550 MHz frequency using LN2  cooling, surpassing his previous achievement with custom liquid cooling by 584 MHz. 

At 5,550 MHz, AMD's Ryzen Threadripper Pro 7995WX scored 186,756 points in Cinebench R23, whereas at 6,000 MHz, the 96-core CPU scored 201,501 points, which is a considerable improvement from the 100,000 points scored at its standard clock speed. 

AMD's Ryzen Threadripper Pro 7995WX operates at a 2.50 GHz base frequency using 350W. Reaching 4.80 GHz with air cooling and 6.0 GHz using liquid nitrogen is a major achievement. After all, if you cool a CPU down using a liquid with a −196 degrees Celsius (−320 degrees Fahrenheit) temperature, you can indeed get a remarkable result. 

Meanwhile, the LN2 overclockers had to adjust memory configurations, reduce the number of DDR5 modules to two (from four) and lower data transfer rates to 3,000 MT/s (from 3,200 MT/s). A common practice in maximizing CPU overclocking potential, but this is something that is never used in real workstations. Speaking of workstations, we can only wonder whether any boutique PC maker attempts to build a factory-overclocked Ryzen Threadripper machine.

Our CPU benchmark hierarchy provides a broad view of relative performance for the latest Intel and AMD processors. Over the last 30 years, Tom’s Hardware has been benchmarking CPUs, and we use the rankings here as the basis of our best CPUs for gaming and best budget CPU rankings. We run over 200 individual tests for each CPU we look at, and that comprehensive performance is condensed here for a high-level view of how CPUs compare across gaming, single-threaded, and multithreaded performance.

Each of our CPU benchmarks helps expose different aspects of performance, from heavily-threaded code compilation and data science workloads to lightly-threaded web apps and audio encoding. We’re currently in the process of the biggest refresh to our CPU benchmarks hierarchy ever, spanning over a decade of processor releases. The results here provide the first half of that testing, focusing on DDR5 platforms that span the AMD vs Intel product lineups. As we fill out our legacy benchmarks, you’ll see more CPUs added to our rankings. If you want to check the performance of older CPUs now, you can use the second page of this article to see our legacy benchmarks.

In games, AMD’s Ryzen 7 9850X3D is the fastest CPU on the market, though other Zen 5 X3D offerings like the Ryzen 9 9950X3D and Ryzen 7 9800X3D aren’t far behind. X3D chips dominate the charts for gaming at 1080p, with the other exception being the relatively unpopular (and expensive) Ryzen 9 7900X3D. Otherwise, Intel’s last-gen Core i9-14900K is the fastest offering from Team Blue, with the new Core Ultra 7 270K Plus coming in slightly behind with Intel’s new iBOT feature.

Intel pulls out strong positions in applications; however, with the Core Ultra 7 270K Plus topping the charts in single-threaded performance and coming in third in multi-threaded rankings. It’s only beaten by the Ryzen 9 9950X and its X3D variant, and only by a hair. Further, both of those CPUs cost about twice as much. AMD's recent Ryzen 9 9950X3D2 claims the top slot in overall performance, but at $900, it's too expensive for most buyers.

In each section below, we’ll show you the rankings for each CPU, as well as reveal what tests went into creating the rankings. We’ll also give you some pointers for benchmarking your own CPU to see how much performance an upgrade or overclock netted you, along with some common, easy-to-run benchmarks you can perform yourself.

CPU Benchmarks Rankings 2026

In the album above, you can see our master charts for gaming, single-threaded, and multi-threaded performance for CPUs. For games, all of our testing was done with an Nvidia RTX 5090 FE, and for applications, our testing was done with an Nvidia RTX 2080 Ti FE. For applications, no compute is actively running on the GPU; it’s a glorified display output that shares a driver with our gaming GPU. You can find a full breakdown of the test benches we used at the end of this article.

Gaming CPU Benchmarks Rankings 2026

You can see the relative score for AMD and Intel CPUs above, measured against the Ryzen 7 9850X3D, which is the fastest gaming CPU on the market, per our testing. So, the Ryzen 7 9800X3D offers 97.04% of the performance of the Ryzen 7 9850X3D, while the Ryzen 9 7900X offers 69.28% of the performance. You can set any CPU as a baseline for comparison with Bench, which is available in Tom’s Hardware Premium.

All of our gaming tests were run with the RTX 5090 FE at 1080p with a mixture of High and Ultra settings. We run each test multiple times — usually between three and five — and pick the median result. In other words, the results we use are real, recorded runs, not an average of several different runs. This is important as some games, such as Far Cry 6, show great CPU scaling but are otherwise inconsistent run-to-run.

In addition to consistent hardware (test benches at the end of this article), we use a consistent test image between platforms. That means the same GPU driver, the same Windows install, the game version, etc. We also tested with Virtualization-Based Security (VBS) turned off, Resizable BAR turned on, and automatic overclocking features disabled. That includes the Intel Extreme power profile and AMD’s PBO, both of which aren’t covered under standard warranty.

For this refresh, we tested 17 games and then calculated a geometric mean of the results. A simple average would provide skewed results with such a large test pool. A geomean provides a more realistic view of how each CPU compares to the others.

Here are the games that we used for testing:

• Counter-Strike 2
• The Last of Us Part One
• Cyberpunk 2077
• Starfield
• A Plague Tale: Requiem
• Hogwarts Legacy
• F1 24
• Marvel’s Spider-Man 2
• Baldur’s Gate 3
• Monster Hunter: Wilds
• Final Fantasy XIV
• Microsoft Flight Simulator 2024
• Doom: The Dark Ages
• Oblivion Remastered
• Far Cry 6
• Hitman 3
• Minecraft RTX

We’re constantly evaluating new games to include in our test suite — see our recent stories on Resident Evil Requiem CPU scaling and Crimson Desert CPU scaling — but we maintain a list of tried-and-true benchmarks for our hierarchy rankings. We want to avoid including brand-new titles, which may see many updates, to keep our rankings as true to reality as possible. If you want more about the rationale behind our game choices, see our behind the scenes look at our CPU hierarchy testing.

Single-Threaded CPU Benchmarks Rankings 2026

We run hundreds of tests for each CPU, but only a small subset of those tests factor into our single-threaded rankings. We use the mp3 encoder LAME with a single thread (both standard and extended), Cinebench 2026 and 2024’s single-threaded test, the ray-traced renderer POV-ray, and WebXRT4, which runs a series of browser-based applications written in various languages.

The fastest chip in the pool here is the Core Ultra 7 270K Plus, which scores 100%, with every other chip scored relative to it. The Core i9-14900K offers 95.4% of the single-threaded performance of the Core Ultra 7 270K Plus, the Ryzen 5 9600X offers 89% of the performance, and so on.

Most real-world workloads aren’t strictly single-threaded, which is why we include it on a subset of the total tests we run. The goal is to see what relative performance looks like in lightly-threaded applications, as well as look into the overall architecture of different CPUs. Single-threaded performance exposes a lot about the architecture in a way that heavily-threaded applications tend to mask.

We’re, of course, looking at performance on a single core, favoring high clock speeds and IPC (instructions per cycle). However, single-threaded performance also says a lot about what’s going on elsewhere inside the CPU, from the speed of the IMC (integrated memory controller) to the fabric/ring speed. That’s why we see things like the Core Ultra 7 270K Plus outperforming the Core Ultra 9 285K, despite the latter sporting higher clock speeds.

Multi-Threaded CPU Benchmarks Rankings 2026

Similar to single-threaded rankings, we use a subset of the total tests we run for CPU reviews in ranking multithreaded performance. Cinebench and POV-ray show up here again, this time using as many threads as possible, alongside VRay, four Blender tests, and Handbrake using various codecs. Although most applications will leverage multiple threads these days, we’re specifically looking at applications that will take as many threads as possible to maximize compute.

Compared to single-threaded workloads, heavily-threaded tasks are less concerned with clock speed and put a greater emphasis on interconnects and core-to-core latency. Core count is obviously important, as well, though it’s been somewhat undermined by Intel’s hybrid architectures over the last several generations.

Given that we’re spanning multiple nodes, core count alone isn’t indicative of higher multithreaded performance. Yes, higher core counts within the same generation will usually provide higher multithreaded performance, but a slew of other factors can increase performance, as well, from all-core and uncore frequencies to higher transistor density. Because of the wide swath of factors, you can see much more aggressive scaling with our multithreaded rankings compared to single-threaded rankings.

Integrated GPU Gaming CPU Benchmarks Rankings 2026

Here's our list of gaming performance with integrated graphics on several of the leading APUs available. We've split this into two different price ranges, so be sure to flip through all of the performance charts. For a bit of commentary and analysis of these results, head to our Ryzen 7 5700G, Ryzen 5 5600G, and Ryzen 3 5300G reviews. The most powerful chip gets a 100, and all others are scored relative to it.

How to Benchmark your CPU

It’s important to know how to benchmark your CPU. It gives you a way to compare performance after an overclock or a CPU upgrade, and it allows you to check if you’re getting the full performance out of your system. Maybe a poor CPU cooler mount is limiting your performance, or maybe your BIOS settings aren’t optimal. Using benchmarks to compare your results lets you see where your rig stacks up, not only for leaderboard purposes, but also basic troubleshooting.

The key to benchmarking your CPU is consistency. The only variable that should change is your CPU, be it a new CPU or an overclock/undervolt. Before starting, make sure to close any applications running in the background. That’s not only to net peak performance, but also to avoid any inconsistencies between runs. Background apps can gobble up threads inconsistently, making it difficult to compare your results from run to run.

If you want more consistency, you can optionally run the following command before benchmarking in an elevated command prompt:

Rundll32.exe advapi32.dll,ProcessIdleTasks

This will force Windows to perform the background tasks it normally does when your PC is idle. It’s not essential, but it’s a good sanity check to make sure there’s nothing interfering with your results.

For applications, you want to test the apps you actually use. If you use the Adobe suite, for example, you can download and use PugetBench for free and compare your results with Puget’s database. A lot of apps don’t have these easy-to-use benchmarking tools and databases, so you need to find a proxy. For instance, Procyon Office measures Microsoft Office performance, but a license costs nearly $1,600 per year. PCMark 10 Basic, which is free, measures open-source office applications. Below, we have some of our favorite free benchmarks for comparing CPU performance.

In games, you can take two approaches: manual or automated. Some modern games include built-in benchmarking tools, such as Cyberpunk 2077 and Doom: The Dark Ages, and although they aren’t perfect, they’re easy to run and highly repeatable. The best way to measure CPU gaming performance, however, is manual benchmarking.

That involves finding a scene where you can go over a specific path repeatedly. That could be starting from a specific checkpoint that you can reload or resorting to a manual save where you start from the exact same position. Regardless, it’s important to avoid randomness in your testing. Keep the path consistent — for example, a walking path through a town — and try not to swing the camera around.

For measuring performance in games, you’ll need a performance monitoring tool. There are simple apps like Nvidia’s FrameView, which logs a ton of information but is a little cumbersome to deal with; it exports data to spreadsheets. CapFrameX is a good alternative, which uses the same backend as FrameView (Intel’s PresentMon), but comes with a user-friendly GUI and extra features like the ability to generate charts right in the app.

After you run your benchmarks, you need a comparison point. Databases like Puget are your best resources on that front. If you’re comparing results to reviews, forum threads, or other systems, keep in mind the variables that can influence performance. It’s not a good idea to compare performance with uncontrolled variables unless you have a wide swath of comparison points.

Best CPU Benchmarks You Can Run

• Cinebench 2026 – Cinebench is the quintessential CPU benchmark, used almost universally in reviews, and it’s completely free to download and use.
• Geekbench 6 – Geekbench has a number of issues, but it offers a massive database for comparing your system against other similar systems. And it’s free to download and run.
• Blender – Blender has a benchmarking utility with a GUI that’s free to download, as well as a large database of results.
• PCMark 10 Basic – The main PCMark 10 benchmark is free to use with the Basic edition, allowing you to test productivity performance with open-source office apps, as well as compare your scores with UL’s database.
• Handbrake – Handbrake is a powerful, free, and open-source video transcoding tool, and it’s easy to run benchmarks with. Use any video file, make sure your settings are the same, and start a stopwatch to measure the time encoding takes. Lower is better.
• WebXPRT 5 – WebXPRT runs a variety of web applications directly in your browser, for free, and with a database to compare results to. It takes a while to run, however.
• JetStream 2 – JetStream is a faster browser-based benchmark, though it doesn’t have a database of results.
• CPU-Z – CPU-Z isn’t a reliable benchmark for real-world performance, but it includes single- and multithreaded tests, it’s easy to run, and you’ll find results online almost as commonly as Cinebench results.
• Y-cruncher – This test calculates Pi with digit extraction, and it’s accelerated with SIMD instructions like AVX. You can only run it from a command line, but it’s relatively straightforward.
• PugetBench – Puget includes benchmarks for the biggest apps in the Adobe suite, as well as DaVinci Resolve. The benchmark itself is free, and Puget maintains a large database. You’ll need a license for the applications it tests, however.

2026 CPU Benchmarks Test System and Configuration

• AMD Ryzen 7 7800X3D vs Intel Core i9-13900K vs Core i7-13700K: Big Gaming Punch, Smaller Price Tag
• AMD Ryzen 9 7950X3D vs Intel Core i9-13900K Faceoff: Battle of the Gaming Flagships
• Intel Core i5-13600K vs AMD Ryzen 7 7700X and Ryzen 5 7600X Face Off
• AMD Ryzen 9 7950X vs Intel Core i9-13900K Face Off
• Ryzen 7 5800X3D vs Core i7-12700K and Core i9-12900K Face-Off: The Rise of 3D V-Cache
• Intel Core i9-12900K vs Ryzen 9 5900X and 5950X: Alder Lake and Ryzen 5000 Face Off
• Intel Core i5-12600K vs AMD Ryzen 5 5600X and 5800X Face Off: Ryzen Has Fallen
• Intel Core i7-12700K vs AMD Ryzen 9 5900X and 5800X Face Off: Intel Rising
• Intel Core i5-12400 vs AMD Ryzen 5 5600X Face-Off: The Gaming Value Showdown

• MORE: Best CPUs for Gaming
• MORE: AMD vs Intel
• MORE: 13th-Gen Raptor Lake All We Know
• MORE: Zen 4 Ryzen 7000 All We Know
• MORE: How to Overclock a CPU
• MORE: How to check CPU Temperature
• MORE: All CPUs Content

2020 - 2022 CPU Benchmarks Hierarchy

You can find our rankings of the most current-gen systems on the previous page. The results below are from our legacy benchmarks, using a different GPU and test systems than our current CPU benchmark rankings. However, this provides great historical context and also includes other previous-gen CPUs not included in our new test suite. You'll also find our even older legacy rankings further below. These date back over the last decade.

Legacy: 2020 - 2022 CPU Benchmarks Rankings - Windows 10 and 11

Legacy: 2020 - 2022 Gaming CPU Benchmarks Ranking

These tests are from our 2022 test bench. We measured performance for the 1080p CPU gaming benchmarks with a geometric mean of Borderlands 3, Hitman 2, Far Cry 5, Project CARS 3, Red Dead Redemption 2, and Shadow of the Tomb Raider. We measured performance for the 1440p CPU gaming benchmarks with a geometric mean of Borderlands 3, Project CARS 3, Far Cry 5, Red Dead Redemption 2, and Shadow of the Tomb Raider. We conducted these tests in Windows 10.

Legacy: 2020 - 2022 Single-Threaded CPU Benchmarks Rankings

These results are from our 2022 test bench. We calculate the above single-threaded CPU benchmark rankings based on a geometric mean of the Cinebench, POV-Ray, and LAME CPU benchmarks. The most powerful chip gets a 100, and all others are scored relative to it. We conducted these tests in Windows 10.

Legacy: 2020 - 2022 Multi-Threaded CPU Benchmarks Rankings