Clean installed Windows 10 22H2 vs Windows 11 23H2 benchmarked for performance

Sayan Sen contributed to this feature, and also provided the benchmark graphics.

As you may be aware, we have been benchmarking different scenarios with the Intel Core i9-14900K sent to us. It provides an ultra-modern test bed to test Microsoft's flagship Windows 11 OS against the older, but more trusted, Windows 10 counterpart.

Tech and gaming enthusiasts will swear blindly that their choice of OS is the best: wherever you look online you will find different results, on the one hand, Windows 10 is better, on the other, Windows 11 should be your choice. It's a hard question to answer with so many different variations of hardware and software being used with test results, so why not try to answer it with some of the best hardware currently at our disposal.

So to put it plainly, which is better, the eight-year-old Windows 10, or on the ongoing development of Windows 11? It is a question we, and other websites, have tried to answer before.

First up, here is a run-down of our test system, you may recognize it from the 14th gen Intel Core i5 and i9 CPUs review we did in October:

Test Hardware

• Intel Core i9-14900K (Amazon|Newegg)
• ASUS ROG STRIX Z790-H GAMING WIFI (BIOS v1402) (Amazon|Newegg)
• ASUS ROG STRIX 850W Gold (Amazon|Newegg)
• Thermalright Peerless Assassin 120 SE ARGB CPU Air Cooler (Amazon|Newegg)
• Carbonaut thermal grizzly 0.2mm (Amazon|Newegg)
• 2x 16GB Corsair Dominator Platinum RGB DDR5 6000MT/s in XMP (Amazon|Newegg)
• Sabrent 1TB Rocket 4 Plus-G (PCIe 4.0) (Amazon|Newegg)
• MSI GeForce RTX 4070 Ti Ventus 3X (Amazon|Newegg)
• Samsung Odyssey G9 Neo at 5120 x 1440 @ 240Hz on DisplayPort (Amazon|Newegg)

Operating System

Windows 10 Pro was clean installed and was updated to the October Patch Tuesday update: build 19045.3570. Then Windows 11 Pro was clean installed to build 22631.2715 (KB5032190), which is the November Patch Tuesday update. On both systems, ASUS Armoury Crate was allowed to install all the drivers that our system needed and Nvidia driver 545.92 was installed for the GeForce RTX 4070 Ti.

The ambient room temperature while testing was measured to be between 19C (66.2 °F) and 20C (68 °F), and our test rig was open plan.

Benchmarks

First up, the GPU bound benchmarks, where the RTX 4070 Ti was pushed to its limit.

In 3DMark, supplied to us by UL Benchmarks, you can see there is barely any difference here between Windows 10 or 11, with the GPU score on 11 edging out Windows 10 ever so slightly, though it is well within the margin of error. The above tests were run three times, and the highest scores were used.

Interestingly, the combined score is higher on Windows 10 despite Windows 11 getting the higher GPU score. Hence, it seems like the processor utilization on the older OS is better. We do gain more insight on this later on in the Time Spy CPU benchmarks section.

While Time Spy is based on the newer DirectX 12 API, Fire Strike Extreme is built on the older DirectX 11. Both Time Spy and Fire Strike Extreme run at 2560 x 1440.

Fire Strike Extreme is a slightly different story as the combined scores are almost identical on the two Windows versions, though once again, Windows 11 manages to edge out Windows 10 by a slim margin. However, just like on Time Spy, the advantage in favor of Windows 11 23H2 is not significant, either statistically or perceptibly.

After testing a couple of synthetic benchmarks, we move on to test a real game.

Shadow of the Tomb Raider's built-in benchmark (released in 2018) does not differ in the stand-off between the two, with Windows 11 again coming out on top (best score after three runs) in all three framerate metrics: average, 95 percentile, and minimum. We start the framerate axis at 60 since 60fps (frames per second) and above is what is considered playable.

From gaming, we move to a bit of productivity testing next. Cinebench 2024 was allowed to run for 10 minutes three times, after which the highest score was used. Here we can see a significant 668 points difference between the older and newer OS. That is a ~3.49% advantage in favor of Windows 11.

If you recall, the new version of Cinebench introduces GPU rendering and uses the RedShift engine. This directly helps us compare CPU and GPU rendering as the rendering scene is the same, unlike on some earlier Cinebench versions.

Hence, across gaming and productivity, it's clear that Nvidia is somehow managing to take better advantage of the newer Windows 11. Although the performance difference is nothing to shout about, this data suggests that Windows 11 is slightly better in the graphics department than Windows 10 under the same testing conditions.

Next up are the CPU-bound benchmarks where we test the might of the new Intel flagship.

As you can see, it's a different story when it comes to CPU performance between the two Windows versions. With Time Spy, Windows 10 edges out Windows 11 in the CPU score by 1,316 points (best score after three runs) and that translates to a difference of 5.83%. That is a serious lead in favor of Windows 10.

It's sort of a similar situation in Fire Strike Extreme though the difference this time is even greater. Here, Windows 10 edges out Windows 11 in Physics by a whopping 6,788 points or ~13.16%. This however does not reflect much in the overall score as Windows 11 is just 2 points behind (best score after three runs).

Looking at these CPU score differentials in the 3DMark test suite does surprise us considering Intel and Microsoft have both confirmed how the Windows 11 is optimized for the hybrid architecture.

Moving on from synthetics, we tested Shadow of the Tomb Raider to see if the same holds true in a real game as well; and indeed, Windows 10 comes out on top, edging Windows 11 ever so slightly in CPU scoring in all of the test scores that were recorded, this was based off the scores of three runs of the benchmark. Once again, we start the framerate axis at 60 since 60fps and above is what is considered playable.

In the CPU Profile test, which tries to gauge how good the scaling is across CPU threads, Windows 10 once again takes the crown, besting Windows 11 at every stage. The highest scores were used after running the benchmark three times. Again, this too is surprising given how Thread Director should make Windows 11 do better on Intel's hybrid CPUs.

The PCMark 10 performance is a mixed bag with Windows 10 and 11 sharing the victories. Overall Windows 11 comes out on top as it wins in two categories: Essentials and Digital Content Creation. The performance lead for Windows 11 in the case of Rendering & Visualisation is especially noteworthy, where it is 6,937 points or 38.5% ahead.

Windows 11 again beat out Windows 10, though, by only 3 points in the Cinebench 2024 CPU benchmark. Interestingly, however, this is nowhere near the 38.5% performance difference we found in the PCMark 10 rendering workload.

We also ran the default 32MB 7-Zip benchmark which measures in GIPS (billion instructions per second). Here Windows 10 edges out Windows 11 in decompression, although Windows 11 performs better in compression, but the differences are once again within the margin of error.

So what have we proven here? It seems like Windows 10, in the case of gaming, is more optimized for CPU while Windows 11 edges out the older OS for GPU performance. Overall, it's generally a close tug-of-war between the two. Microsoft would have a hard time claiming much of a significant performance jump between the older and newer Windows.

Would it convince me to permanently switch to Windows 11? Not in its current state, nope. But despite this, there is no evidence based on these tests that Windows 11 is worse in gaming and productivity. You could probably safely just go with what you prefer and not lose a bunch of gaming or productivity performance, as our tests prove that they are mostly marginal.

Check out our other benchmark coverage of Windows 10 clean install vs Windows 11 in-place upgrade and our Intel Core i5-14600K and i9-14900K review.

As an Amazon Associate I earn from qualifying purchases.