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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 279621 | M3 Ultra (60c) | 256 GB | Qwen3.5-122B-A10B | 4bit | 4k | 738.0 | 52.0 | 26-03-19 |
| 279622 | M3 Ultra (60c) | 256 GB | Qwen3.5-122B-A10B | 4bit | 8k | 722.6 | 48.8 | 26-03-19 |
| 279623 | M3 Ultra (60c) | 256 GB | Qwen3.5-122B-A10B | 4bit | 1k | 711.6 | 54.8 | 26-03-19 |
| 279624 | M4 (10c) | 16 GB | Qwen3.5-4B | 4bit | 1k | 410.7 | 38.4 | 26-03-19 |
| 279625 | M2 Ultra (60c) | 64 GB | QwQ-32B | 4bit | 4k | 241.6 | 26.8 | 26-03-19 |
| 279626 | M2 Ultra (60c) | 64 GB | QwQ-32B | 4bit | 1k | 240.3 | 29.7 | 26-03-19 |
| 279627 | M4 Pro (16c) | 24 GB | gpt-oss-20b-MXFP4-Q8 | 4bit | 4k | 643.7 | 61.0 | 26-03-19 |
| 279628 | M4 Max (32c) | 36 GB | Qwen3.5-35B-A3B | 4bit | 4k | 1,423 | 91.2 | 26-03-19 |
| 279629 | M3 Ultra (60c) | 256 GB | Qwen3.5-35B-A3B | 4bit | 16k | 1,940 | 77.7 | 26-03-19 |
| 279630 | M4 Pro (16c) | 24 GB | gpt-oss-20b-MXFP4-Q8 | 4bit | 64k | 373.5 | 21.1 | 26-03-19 |