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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 156591 | M3 Ultra (80c) | 512 GB | MiniMax-M2.7 | 6bit | 8k | 701.5 | 33.4 | 26-05-03 |
| 156592 | M3 Ultra (80c) | 512 GB | MiniMax-M2.7 | 6bit | 4k | 754.1 | 37.4 | 26-05-03 |
| 156593 | M3 Ultra (80c) | 512 GB | MiniMax-M2.7 | 6bit | 1k | 636.3 | 40.9 | 26-05-03 |
| 156594 | M4 Max (32c) | 36 GB | Hermes-3-Llama-3.1-8B | 4bit | 1k | 726.4 | 71.7 | 26-05-03 |
| 156595 | M4 Max (32c) | 36 GB | Hermes-3-Llama-3.1-8B | 4bit | 4k | 718.1 | 64.9 | 26-05-03 |
| 156596 | M5 Max (40c) | 128 GB | Qwen3.6-35B-A3B-UD | 4bit | 4k | 3,837 | 111.4 | 26-05-03 |
| 156597 | M5 Max (40c) | 128 GB | Qwen3.6-35B-A3B-UD | 4bit | 1k | 2,335 | 115.1 | 26-05-03 |
| 156598 | M5 Max (40c) | 128 GB | Qwen3.6-35B-A3B-UD | 4bit | 1k | 2,351 | 117.0 | 26-05-03 |
| 156599 | M5 Max (40c) | 128 GB | Qwen3.6-35B-A3B-UD | 4bit | 4k | 3,038 | 112.9 | 26-05-03 |
| 156600 | M5 Max (40c) | 128 GB | Qwen3.6-35B | 8bit | 128k | 1,398 | 56.0 | 26-05-03 |