AMD says that gaming performance on Ryzen 9000 is actually better than what you’ve read. As you can read in our Ryzen 9 9950X and Ryzen 9 9900X review, AMD’s new Zen 5 CPUs are the best processors you can buy when it comes to productivity. Gaming performance, on the other hand, is disappointing.
According to a new blog post from AMD, there are a few reasons why reviewers saw lower gaming performance than expected. Chief among them are the fact that AMD used an unreleased version of Windows 11 — the 24H2 update, which is available to Windows Insiders — and that it used an administrator account for its “automated test methodology.” In light of that, I downloaded the Windows update, spun up an admin account, and retested the Ryzen 9 9950X. And I’m not seeing what AMD claims at all.
| Original Ryzen 9 9950X | 24H2 Admin Ryzen 9 9950X | Percent change | |
| Ashes of the Singularity | 59.8 frames per second (fps) | 57.4 fps | -4% |
| F1 2022 | 346.7 fps | 351.4 fps | +1.4% |
| Far Cry 6 | 177.5 fps | 179.7 fps | +1.2% |
| Final Fantasy 14 Dawntrail | 224.8 fps | 241.1 fps | +7.3% |
| Hitman 3 | 201.8 fps | 203.1 fps | +0.6% |
| Tiny Tina’s Wonderlands | 274.4 fps | 274.3 fps | 0% |
Get your weekly teardown of the tech behind PC gaming
The only real difference I saw was in Final Fantasy 14: Dawntrail. Even more concerning is that I saw basically no difference in Far Cry 6, and AMD claims that it saw a 13% increase with these updates. It’s been weeks now of second-guessing my own results, reinstalling Windows countless times, and now setting up AMD’s ideal test configuration to somehow, maybe, produce a result that lines up with what it claims. I’m just not seeing it.
That’s ignoring the obvious issue with AMD’s claims here. It’s using a Windows 11 version that isn’t available widely now, and a separate administrator account. Keep in mind that this account isn’t just your Windows account with administrator privileges — it’s an entirely separate account on your PC. In addition, these changes to Windows don’t just benefit Zen 5. They also benefit Zen 4 and Zen 3 CPUs, essentially making the performance uplift null.
So, where did AMD get its claims about leadership in gaming performance against both its last-gen CPUs and the competition from Intel? Against last-gen options, AMD’s review guide provides a lot of hints. There, AMD shows the Ryzen 9 9950X leading the Ryzen 9 7950X by 26% in League of Legends, despite the fact that the gains in other games are in the low single-digits. Worse, AMD tested a version of League of Legends that’s three years old. It tested 11.5. The current version is 14.16. A data point like this only serves to inaccurately inflate the average.
Against Intel, AMD says in its blog post that it used “Intel default settings-baseline power profile” for testing, which is as clear as mud. We’ve seen the instability fiasco with Intel’s latest CPUs, as well as how its default power settings can impact gaming performance by upwards of 9%. In addition, Intel explicitly said not to use the baseline power profile on motherboards capable of higher values. I understand AMD wants to paint its products in the best light, but testing with the baseline power settings teeters on the edge of misleading.
Since the launch of Zen 5 CPUs, there has been a lot of digging to uncover some hidden gaming performance within the CPUs. Maybe it’s there, but I haven’t seen it yet. Like any new architecture, AMD has some growing pains with Zen 5, but that doesn’t change my recommendation on the CPUs. Until you can buy one of these processors, throw it in a motherboard, and get the performance AMD claims, I wouldn’t buy one. The CPUs aren’t there yet.
