Core Ultra 9 275HX beats AMD's flagship Ryzen 9 mobile chip by 7% in PassMark — 34% faster than the i9-14900HX
Let's see how it holds up against AMD's upcoming Ryzen 9000HX3D chips.

Preliminary benchmarks at PassMark put Intel's new Arrow Lake-based Core Ultra 9 275HX 7% ahead of AMD's Ryzen 9 7945HX3D processor in CPU Mark. Moreover, despite a 400 MHz clock deficit, the 275HX manages to land 34% faster than last generation's i9-14900HX, with a solid 9% lead when looking at single-threaded performance. Because this test is based on only one sample of the processor, it is suggested to wait for in-depth, independent reviews before jumping to conclusions. Gaming laptops equipped with Intel's Arrow Lake-HX family are expected to launch soon, most likely hitting shelves next month with the launch of Nvidia's RTX 50 mobile series.
Arrow Lake-HX CPUs act as desktop replacements, employing an architecture similar to Core Ultra 200S chips on desktop, but in a BGA package. There are some minor differences between these CPUs and Arrow Lake-H, including the elevated TDPs and the lack of LPE cores on the SoC tile. These are Intel's highest-performing laptop chips this year, designed to be paired with a dedicated GPU, with a PBP (Processor Base Power) of 55W, going as high as 160W depending on the use case.
The Core Ultra 9 275HX in question is only second to the flagship 285HX, both offering 24 cores (8P + 16E) and 24 threads. The 275HX clocks at a maximum of 5.5 GHz on the performance cores, a 400 MHz deficit versus the last generation, and 100 MHz slower than the 285HX. Nonetheless, Intel has managed to extract somewhat decent efficiency and performance gains thanks to the newer TSMC N3B process and updated Lion Cove and Skymont architectures.
In the single-threaded benchmark at PassMark, the 275HX leads the pack; up to 16% faster than the Ryzen 9 7945HX3D. This is not an exact apples-to-apples comparison as AMD's new Zen 5-based Fire Range chips are just around the corner, so we aren't sure if Intel can maintain its performance throne for long. In any case, this lead extends to 34% against the i9-14900HX in CPU Mark; PassMark's proprietary tool for estimating a CPU's performance. Despite a near 7% reduction in boost clocks, the Arrow Lake chip still lands roughly 9% ahead in terms of single-core performance.
The Core Ultra 9 285H (Arrow Lake-H based) delivered lackluster efficiency numbers in a recent user review. While this could simply boil down to the dated Meteor Lake SoC tile, which shouldn't carry over to Arrow Lake-HX, it's best to hold off until detailed reviews are available to get a clearer picture.
Given the TDP of these chips, performance will largely depend on the laptop's thermals and power delivery capability. The integrated GPU has been updated to Alchemist+ with support for XMX supporting DPAS instructions. None of Intel's Arrow Lake chips are compliant with Microsoft's CoPilot+ requirements, except for Lunar Lake. Laptops packed with Intel's Arrow Lake-HX chips are slated for a late Q3 launch, so you'll probably see a handful of these on shelves next month.
Stay On the Cutting Edge: Get the Tom's Hardware Newsletter
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Hassam Nasir is a die-hard hardware enthusiast with years of experience as a tech editor and writer, focusing on detailed CPU comparisons and general hardware news. When he’s not working, you’ll find him bending tubes for his ever-evolving custom water-loop gaming rig or benchmarking the latest CPUs and GPUs just for fun.
-
KyaraM
Of course it's just advertising and bias when Intel beats AMD. Fact not withstanding that the Ryzen has 32 threads while the Intel chip has only 24 since it lacks hyperthreading. Sometimes I wonder if some people here even think about what they write, or if a short circuit happens in their brain and they jump immediately to "WAAAAAH, IMPOSSIBLE, FAKE, UNFAIR COMPARISON!!!" the second an Intel chip lands before an AMD chip in any one bloody benchmark.das_stig said:<<Removed by moderator>> -
-Fran- This is the key element missing here: "Given the TDP of these chips, performance will largely depend on the laptop's thermals and power delivery capability".Reply
Without this information, may as well be comparing TDP-configured 45W 4070s to 150W 4060s and say the 4060 is the superior dGPU in laptops.
Still, better than having a nasty surprise like in the desktop. At least, so far, ArrowLake on mobile won't suck? We hope?
Regards. -
bit_user
IMO, that qualifies as an ad hominem attack. Instead, I think you should present your evidence and let people decide for themselves. If you're so sure he's regurgitating copy from Intel's marketing department, surely you have some evidence for that?das_stig said:<<Removed by moderator>>
If you're concerned that he's using these benchmark leaks to spin a misleading narrative, perhaps based on some problem with PassMark, why not just say so? Absent a problem with the data cited, I think you don't really have a case here. -
A Stoner Comments are raging here.Reply
I came here to comment that if Intel is doing well here, that is great news for everyone. Lately I have been fanboying for AMD because they are the ones bringing more innovation to the masses than Intel has.
The reality is that I do not care who brings the wins, but I do want both sides bringing punches and landing hits because that is what moves technology forward for us all.
They put the caveats down. This is one processor in one set of tests, and we do not know all the details. -
heffeque This "kingdom" might be short-lived, as the 9000HX is around the corner.Reply
But even if the 9000HX doesn't surpass this chip... who's going to trust it to not burn like the latest Intel chips? -
Mr Majestyk I'm fully an AMD user, but I doubt 9955HX3D will help much. Zen 5 has been a pitiful update, at least on Windows. Arrow Lake really only sucks at gaming.Reply
Frankly, I'm waiting for Panther Lake before I finally update my dinosaur 2016 laptop. I don't think AMD has anything new this year outside of Halo, but it will chug a lot more power than Panther Lake. For AMD I'll wait for Medusa Point/Halo. -
Quirkz
Absolutely correct. Especially when the article says these chips will burst up to 160W - that's desktop power draw, not laptop.-Fran- said:This is the key element missing here: "Given the TDP of these chips, performance will largely depend on the laptop's thermals and power delivery capability".
Regards.
The critical laptop criteria here is: What's the performance difference at the same power draw? (especially a meaningful laptop class power draw of around the base 55W - which is already high enough to mean less than two hours battery life!)
Either way, I welcome Intel returning to competitive form. That's good for us consumers: Look at the raw CPU power of laptops now, compared to the 4 core standard laptop of 5 years ago. -
KyaraM
I mean, a 7945HX can easily draw 130W+ during burst, maybe even more with the "right" settings; it's not just Intel having high maximum burst power draw. My newest system has that chip installed and definitely gets there, at least for around 4 minutes before it hits the TAU limit, or whatever AMD calls it. It's on an mITX board (the Minisforum BD795i SE, to be precise), though, so cooling is a lot better than it would be in a laptop or mini PC.Quirkz said:Absolutely correct. Especially when the article says these chips will burst up to 160W - that's desktop power draw, not laptop.
The critical laptop criteria here is: What's the performance difference at the same power draw? (especially a meaningful laptop class power draw of around the base 55W - which is already high enough to mean less than two hours battery life!)
Either way, I welcome Intel returning to competitive form. That's good for us consumers: Look at the raw CPU power of laptops now, compared to the 4 core standard laptop of 5 years ago.
A 55W comparison would be interesting, though. Not because of battery life, that was never a good argument to me; if you are in a place where you have no access to a wall plug, you are most likely also in a place where you shouldn't work to begin with, and same for gaming. It's because of heat dissipation. I know how hot my small 7640HS in the gaming laptop can get at 54W; I don't want to know how hot these chips get north of 100W. -
mitch074
Zen 5, "pitiful" ? It still delivered +10% IPC improvements while retaining similar or better thermals, when Intel brought... Lower performance with, admittedly, much better thermals than before.Mr Majestyk said:I'm fully an AMD user, but I doubt 9955HX3D will help much. Zen 5 has been a pitiful update, at least on Windows. Arrow Lake really only sucks at gaming.
Frankly, I'm waiting for Panther Lake before I finally update my dinosaur 2016 laptop. I don't think AMD has anything new this year outside of Halo, but it will chug a lot more power than Panther Lake. For AMD I'll wait for Medusa Point/Halo.
It is telling when the only way Intel can deliver a better product is when the OS is tailored for it and nerfing the competition. The surge of interest for SteamOS (which is Linux based) is proof that Wintel isn't exactly unassailable anymore.
Not that I don't appreciate some healthy competition, AMD are resting a bit on their laurels these days. -
Greg In Bothell "Because this test is based on only one sample of the processor..."Reply
Apparently Tom's editorial department has no one at the wheel to stop bad stories like this from getting published.
The sample size makes this number a huge nothingburger and it's massively irresponsible to hype it. I've seen it in recent days on Passmark's charts and dismissed it as being too untrustworthy.
When you wrote about it, I thought maybe more samples had come in and there was more credibility. Nope. You just glommed onto it for clickbait, then provided a mild caveat in the body of the article. That's bad journalism.