Intel Core Ultra 9 285K vs AMD Ryzen 7 9800X3D
Processor comparison for gaming — benchmark scores, estimated framerates and which one suits your build.
The short answer
AMD Ryzen 7 9800X3D is the faster of the two for gaming, by roughly 8% over the Intel Core Ultra 9 285K.
Gap measured from estimated framerates across 5 games, not from benchmark scores — scores rank parts reliably but compress the distance between two fast ones.
Note on 3D V-Cache: the AMD Ryzen 7 9800X3D carries AMD's stacked cache, which no PassMark figure can see — you can see it in the L3 cache row below. Our gaming score applies a measured premium for it, so the ranking here deliberately differs from a raw multi-core comparison.
Benchmark scores
| Measure | Intel Core Ultra 9 285K | AMD Ryzen 7 9800X3D |
|---|---|---|
| Gaming score ours | 37,721 | 45,886 |
| CPU Mark | 67,256 | 39,950 |
| Faster than | 100% of 1,255 | 97% of 1,255 |
| Cores / threads | 24 / 24 | 8 / 16 |
| Base clock | — | 4.7 GHz |
| Boost clock | — | 5.2 GHz |
| L3 cache | 36 MB | 96 MB |
| TDP | 125 W | 120 W |
| Socket | LGA 1851 | AM5 |
| Architecture | Arrow Lake-S | Zen 5 |
| Process | 3 nm | 4 nm |
| Memory support | DDR5 | DDR5 |
| Integrated graphics | Arc Xe-LPG Graphics (4 Xe-cores) | Radeon Graphics (2 CUs RDNA 2) |
| Launch price | $589 | $479 |
| Released | 2024-10-24 | 2024-11-07 |
The gaming score damps high-core-count parts, which win productivity benchmarks without delivering proportional framerate, and credits 3D V-Cache where it is measurably earned.
Estimated FPS at 1080p
Both measured against the same reference graphics card — a GeForce RTX 4090 — so the processor is the variable being tested. Ranges, not exact figures.
| Game | Intel Core Ultra 9 285K | AMD Ryzen 7 9800X3D |
|---|---|---|
| Fortnite | 382–637 CPU | 387–645 CPU |
| CS:GO | 442–736 CPU | 484–807 CPU |
| GTA V | 151–252 CPU | 165–275 CPU |
| Apex Legends | 147–246 CPU | 175–291 CPU |
| PUBG | 173–288 CPU | 192–321 CPU |
Which suits your build?
These figures hold a reference graphics card constant. Your own build decides the answer: a faster processor buys nothing if the other component is already the limit. Run both against your actual hardware:
Check both in the bottleneck calculator