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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 156381 | M2 (10c) | 16 GB | Qwen3.5-9B | 8bit | 4k | 109.5 | 0.9 | 26-05-02 |
| 156382 | M4 Max (40c) | 128 GB | gpt-oss-20b-MXFP4-Q8 | 4bit | 4k | 1,448 | 105.2 | 26-05-02 |
| 156383 | M4 Max (40c) | 128 GB | gpt-oss-20b-MXFP4-Q8 | 4bit | 1k | 965.7 | 111.1 | 26-05-02 |
| 156384 | M4 Max (40c) | 128 GB | Qwen3-Coder-Next | 4bit | 4k | 1,331 | 80.5 | 26-05-02 |
| 156385 | M4 Max (40c) | 128 GB | Qwen3-Coder-Next | 4bit | 1k | 1,059 | 85.5 | 26-05-02 |
| 156386 | M4 Max (40c) | 128 GB | Qwen3.5-27B | 8bit | 4k | 198.3 | 14.0 | 26-05-02 |
| 156387 | M4 Max (40c) | 128 GB | Qwen3.5-27B | 8bit | 1k | 235.0 | 16.2 | 26-05-02 |
| 156388 | M4 Pro (20c) | 64 GB | Llama-3.3-70B-Instruct | 3bit | 1k | 50.5 | 7.5 | 26-05-02 |
| 156389 | M4 Pro (20c) | 64 GB | Llama-3.3-70B-Instruct | 3bit | 4k | 45.3 | 7.0 | 26-05-02 |
| 156390 | M4 Pro (16c) | 64 GB | Qwen3.6-35B-A3B-UD | 4bit | 32k | 585.6 | 48.1 | 26-05-02 |