Find the best hardware for running AI locally
Not just can this model fit, but: how fast will it actually go, what will it cost you to own after tax and electricity, and is there a better option — including not buying anything at all.
Start here
Hardware configurations
21
10 products
Open-weight models
27
96 quantized variants
Measured benchmarks
15
every one traced to a source
Estimated pairings
1,163
labelled, never precise
The rule this site is built on: never hide uncertainty. Measured numbers are precise. Estimates are shown as ranges and marked Estimated. Assumptions are marked Assumption and you can change them. Manufacturer figures are marked Official spec because a theoretical bandwidth is not a measurement. Where four sources disagree, we show the spread instead of an average.
Popular hardwareAll 21 →
Quad RTX 5090 workstation (4x 32 GB)€13.999
128 GB · 7,168 GB/s · 2,250 W
Dual RTX 5090 workstation (2x 32 GB)€6.499
64 GB · 3,584 GB/s · 1,180 W
Dual used RTX 4090 workstation (48 GB)€4.099
48 GB · 2,016 GB/s · 980 W
RTX PRO 6000 Blackwell workstation (96 GB)€11.499
96 GB · 1,792 GB/s · 660 W
RTX PRO 6000 Max-Q workstation (96 GB, 300 W)€11.999
96 GB · 1,792 GB/s · 390 W
RTX 5090 workstation (1x 32 GB)€3.699
32 GB · 1,792 GB/s · 640 W
Popular modelsAll 27 →
Recent measured benchmarksAll benchmarks →
| Hardware | Model | Quant | Context | Decode | Prefill | Type | Date |
|---|---|---|---|---|---|---|---|
| Framework Desktop (Ryzen AI Max+ 395, 128 GB) | gpt-oss-120b | MXFP4 | 4K | 34.0 t/s | — | Measured (community) | 11 May 2026 |
| GMKtec EVO-X2 (Ryzen AI Max+ 395, 128 GB) | gpt-oss-120b | MXFP4 | 4K | 32.0 t/s | — | Measured (community) | 20 Apr 2026 |
| Framework Desktop (Ryzen AI Max+ 395, 128 GB) | gpt-oss-120b | MXFP4 | 4K | 31.0 t/s | — | Measured (community) | 2 Apr 2026 |
| Framework Desktop (Ryzen AI Max+ 395, 128 GB) | gpt-oss-120b | MXFP4 | 4K | 55.0 t/s | — | Measured (community) | 15 Mar 2026 |
| Framework Desktop (Ryzen AI Max+ 395, 128 GB) | gpt-oss-120b | MXFP4 | 4K | 30.0 t/s | — | Measured (community) | 10 Feb 2026 |
| Framework Desktop (Ryzen AI Max+ 395, 128 GB) | Qwen3.6 27B | Q4_K_M | 4K | 11.5 t/s | — | Measured (community) | 10 Feb 2026 |
| Framework Desktop (Ryzen AI Max+ 395, 128 GB) | Qwen3-Coder 30B-A3B | Q4_K_M | 4K | 52.0 t/s | — | Measured (community) | 10 Feb 2026 |
| RTX 5090 workstation (1x 32 GB) | Qwen3 30B-A3B | Q4_K_M | 0 | 352.1 t/s | — | Measured (community) | 30 Nov 2025 |
Popular comparisonsAll →
NVIDIA DGX Spark 128 GB vs Mac Studio M5 Max 128 GBNVIDIA DGX Spark 128 GB vs Framework Desktop (Ryzen AI Max+ 395, 128 GB)Mac Studio M5 Max 128 GB vs Framework Desktop (Ryzen AI Max+ 395, 128 GB)Mac Studio M5 Ultra 256 GB vs RTX PRO 6000 Blackwell workstation (96 GB)RTX 5090 workstation (1x 32 GB) vs Mac Studio M5 Max 64 GBDual RTX 5090 workstation (2x 32 GB) vs RTX PRO 6000 Blackwell workstation (96 GB)Mac Studio M5 Ultra 512 GB vs Mac Studio M3 Ultra 512 GBRTX 5090 workstation (1x 32 GB) vs Used RTX 4090 workstation (24 GB)