Community Benchmarks
Download oMLX →Real performance numbers from oMLX users around the world. Submit yours from the app (v0.2.6+).
| # | Chip | RAM | Model | Quant | Ctx | PP tok/s | TG tok/s | Date ↓ |
|---|---|---|---|---|---|---|---|---|
| 12011 | M2 Ultra (76c) | 128 GB | gpt-oss-120b-MXFP4-Q8 | 4bit | 1k | 314.1 | 67.7 | 26-03-08 |
| 12012 | M4 (10c) | 16 GB | Qwen3.5-9B | 6bit | 1k | 217.2 | 15.7 | 26-03-08 |
| 12013 | M4 (10c) | 16 GB | Qwen3.5-9B | 6bit | 1k | 217.2 | 15.8 | 26-03-08 |
| 12014 | M2 Ultra (76c) | 128 GB | MiniMax-M2.5 | 3bit | 16k | 282.6 | 22.4 | 26-03-08 |
| 12015 | M2 Ultra (76c) | 128 GB | MiniMax-M2.5 | 3bit | 8k | 320.5 | 29.5 | 26-03-08 |
| 12016 | M2 Ultra (76c) | 128 GB | MiniMax-M2.5 | 3bit | 4k | 337.6 | 35.7 | 26-03-08 |
| 12017 | M2 Ultra (76c) | 128 GB | MiniMax-M2.5 | 3bit | 1k | 299.7 | 42.1 | 26-03-08 |
| 12018 | M3 Ultra (60c) | 256 GB | Qwen3-Coder-Next | 8bit | 64k | 994.8 | 34.7 | 26-03-08 |
| 12019 | M3 Ultra (60c) | 256 GB | Qwen3-Coder-Next | 8bit | 16k | 1,434 | 52.0 | 26-03-08 |
| 12020 | M3 Ultra (60c) | 256 GB | Qwen3-Coder-Next | 8bit | 32k | 1,262 | 45.8 | 26-03-08 |