Our evaluations demonstrate that Intel’s solution for its Arrow Lake processors is ineffective in remedying the chips’ mediocre gaming performance, at least based on the motherboards we analyzed. We discovered that the updated gaming performance of the Core Ultra 9 285K with one motherboard is now marginally slower than it was before. Furthermore, the necessary operating system update has enhanced gaming performance for the previous-gen Raptor Lake Refresh even more so than the Arrow Lake processors, leading to the flagship Core Ultra 9 285K lagging even further behind its forerunner. As illustrated in our benchmarks below, the Core Ultra 9 285K continues to fall short of Intel’s original gaming performance marketing assertions and will not be included in our list of top gaming CPUs.
The launch of the Intel ‘Arrow Lake’ Core Ultra 200S was tainted by lower gaming performance than the company had promised, failing to rival the previous-gen Core i9-14900K flagship in gaming despite the company’s already lackluster claims of equivalence. Intel cited a multitude of issues as the cause of the disappointing gaming performance and provided solutions through both Windows and BIOS updates.
At CES 2025, Intel shared its own benchmarks with assertions of up to 25% improved gaming performance with the ‘fix,’ a statement we have unfortunately seen echoed by the media. However, those increases are applicable only to certain very specific issues that not all reviewers and users will have encountered. As we will elaborate on extensively below, there exists considerable ambiguity in Intel’s questionable claims of performance improvements, yet our tests reveal that the updates yield no significant enhancement in the competitive landscape against AMD, and even more troubling, Arrow Lake now performs worse in gaming compared to its predecessor.
Testing Details
Intel’s ‘fix’ necessitates two basic components: Windows 11 build 26100.2314 (or later) and microcode version 0x114 with CSME firmware kit 19.0.0.1854v2.2 (or more recent). For our initial review, we worked with the then-current Windows version 26100.2033. We transitioned to version 26100.2605 for the configurations that demonstrate patched performance. Intel claims that the updated version includes an enhanced Power Performance Management (PPM) package that guarantees optimal performance when using the balanced power profile. However, similar to all reviewers adhering to the most fundamental best practices, we initially tested with the High Performance power profile for Intel systems and thus expect minimal to no impact on our findings from this alteration. You can read about the remaining fixes here.
Cyberpunk 2077 experienced a notable performance boost from a fix released for the game’s code. However, Intel states this was a problem created by the game developers themselves, and the developers rectified the issue independently. Intel further asserts that we should not anticipate additional game code updates that will enhance Arrow Lake’s performance in the future.
Several game titles have likewise received updates since our original assessment, making it unfeasible to replicate those results with the updated test configurations. To maintain consistency, we revisited our original test setup with the launch version of the BIOS and Windows through our test suite to ensure that we do not attribute performance enhancements from game code updates to an ‘fix’ from Intel. Subsequently, we evaluated the effects following updates to the latest BIOS/Windows versions.
Consequently, the entries listed below marked with ‘Original’ signify the original BIOS and firmware but now come with updated testing to represent the present state of the game code. The entries labeled as ‘New FW-OS’ demonstrate testing incorporating the cumulative effects of all updates. For simplicity, we solely evaluated with standard DDR5 memory (no CUDIMMS) and two motherboard platforms.
Receive Tom’s Hardware’s latest news and thorough reviews, directly to your inbox.
Core Ultra 9 285K gaming performance
Image 1 of 14
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
“`html
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)“““html
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
“`
As illustrated above, the Asus motherboard combined with the Core 9 285K indeed experiences a minor decrease in gaming performance post-patch – the unpatched 285K setup is 3% slower than the recently patched version. I repeatedly tested this scenario, and Asus has not yet addressed our inquiries regarding this issue.
We redirected our focus to assessments on the MSI motherboard to ascertain if we should anticipate performance declines across all motherboards. The MSI motherboard commenced from a notably lower baseline with the initial firmware/OS, but it did achieve at least a respectable 3.7% improvement. Nevertheless, it still lags behind the original unpatched Asus configuration with the identical setup we utilized for our evaluation by 1.9%.
More significantly, in comparison to the fastest patched 285K outcomes on the MSI motherboard, the Ryzen 9 9950X now outperforms it by 6.5% (it was approximately 3% quicker in our initial evaluation), and the Ryzen 7 9800X3D continues to be nearly 40% faster than the 285K – the gap is substantial. This suggests the fix has not positively impacted Arrow Lake’s competitive standing compared to AMD’s chips.
Even more worrying for Intel is that its preceding-gen Core i9-14900K saw a much more substantial boost than the Core 9 285K following the update to the new Windows version. We merely updated the OS for the revised 14900K configuration – no new firmware had been provided for our test motherboard since the 285K assessment. As shown, the 14900K is now 7% faster than the evaluations conducted with the older Windows version. It seems that Windows has rectified some issue affecting all Intel processors here, resulting in the 14900K now being 14% faster than the 285K.
For context, we initially recorded the 14900K as being 6.4% quicker than the 285K in our launch day review, but currently, the 14900K stands at 14% faster than the updated 285K. Once more, this falls short of Intel’s original performance assertions regarding the 285K maintaining parity with the 14900K.
Thus far in our gaming performance evaluations and the analyses conducted by other media outlets, while Intel may have resolved a few edge cases, it certainly has not rectified the turmoil generated by setting expectations for the Core Ultra 9 285K unrealistically high. The 285K still fails to meet those standards, and the reality is that the previous-gen Intel processors demonstrably outperform it in gaming.
Core Ultra 9 285K productivity performance
Image 1 of 18
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
“`html
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
“““html
class=”item__image-cont”>
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
“““html
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
That’s not to imply Arrow Lake is entirely unfavorable, of course. The 285K possesses its appeal in generational advancements regarding single- and multi-threaded tasks, although AMD still retains the crown in the latter segment.
Here, it’s evident that the Core Ultra ‘adjustments’ had no effect on the overall performance in our aggregate evaluations of single- or multi-threaded productivity tasks. We did observe a slight decline for the 14900K with the updated operating system, but merely by a minimal 0.8%, which is within the anticipated variance.
Conclusion
Image 1 of 16
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
“““html
(Image attribution: Tom’s Hardware)
(Image attribution: Tom’s Hardware)
(Image attribution: Tom’s Hardware)
(Image attribution: Tom’s Hardware)
(Image attribution: Tom’s Hardware)
“““html
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
“`
(Image credit: Tom’s Hardware)
(Image credit: Tom’s Hardware)
Presented here are Intel’s performance assertions made during a presentation at CES 2025 that details the effects of its update. A number of these matters would not have affected proficient reviewers, for instance, by employing a balanced power profile or failing to confirm that APO was operational (APO influences a limited selection of games that most proficient reviewers examine anyway). Regarding APO, this was available to evaluators for launch-day assessments – which we utilized – and simply required one to affirm it was functional. One of the other instances denotes an enhancement in 7zip, yet that has no correlation to the inadequacies in gaming.
You will also observe that Intel includes enhanced performance in Cyberpunk 2077 in the summary slide. Nevertheless, the company has also mentioned that the concern with that title was self-inflicted by the developers and resolved without Intel’s intervention. Yet it is recorded as an Intel achievement from the ‘remedy.’ Intel does not mention that the 14900K also received a significant boost from the updated Cyberpunk 2077 game code. The same can be stated about Far Cry 6 – the 14900K greatly benefited more than the 285K from the transition to the new version of Windows.
Intel’s presentation distinctly illustrates the performance impact of various features being toggled on or off, yet this can be misleading, and Intel’s assertions are ambiguous and indistinct. You may or may not endure any or several of these challenges, and the effect of each issue could differ significantly based on your configuration. All of the claims in the charts above, or none of the claims, or anything in between, could or could not pertain to you and your system. Intel also asserts that all of these matters may impact you in one instance but not in another, despite no changes (like settings) having been made. Here’s how Intel articulates it:
“The precise performance increase you will encounter with these updates relies on the particular issue, or combination of issues, present on your system when your data was initially collected. Results also depend on your selection of games or applications. Some issues are more elusive than others, are more pertinent to certain workload characteristics, and/or occur intermittently.”
This makes it nearly impossible to contest the claims definitively. Regardless, Intel’s assertions regarding performance being faster or slower with features toggled on or off still do not clarify the real issue – the 285K’s competitive positioning. You’ll note that none of those tests mentioned above include competing processors, be they a previous-generation Intel component or any AMD part. This explains why you do not see significant factors such as the 14900K’s improved performance in Cyberpunk 2077 and Far Cry 6.
These strategies and the testing results render this entire ‘fix’ endeavor seem more like spin rather than a genuine remedy. Yes, Intel acknowledges that it failed to ensure that some issues were rectified in a manner that would apply uniformly across all users, and it has now addressed those concerns. However, it still hasn’t elevated the chips to the level of performance it initially promised, and even if it simply matched its previous-generation chips in gaming as it claimed, that remains unsatisfactory. We anticipate generational improvements in performance, and anything less is rightly criticized.
Ultimately, Intel’s remedies for its various shortcomings did not clearly ‘fix’ the Core Ultra 9 285K’s gaming performance in any appreciable way, and they certainly fall short of meeting the company’s initial marketing assertions or altering the competitive positioning of its subpar Arrow Lake chips. In fact, it appears that Arrow Lake is progressing backward. Despite its other favorable attributes, the Core Ultra 285K simply isn’t the premier choice for gaming.
MSI MPG X870E Carbon WiFi — Games (all 9000 Series with 1.2.0.2a, 7950X3D and 7900X3D 1.2.0.2)
RAM
G.Skill Trident Z5 Neo DDR5-6000
Cooling
Corsair iCue Link H150i RGB
Note:
Microsoft recommends gamers to deactivate certain security features to enhance gaming performance. Consequently, we turned off secure boot, virtualization support, and fTPM/PTT.