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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 206371 | M5 Max (40c) | 128 GB | gemma-3-27b-it-qat | 4bit | 1k | 699.8 | 30.4 | 26-04-16 |
| 206372 | M1 Pro (16c) | 32 GB | gemma-4-26b-a4b-it | 4bit | 1k | 232.6 | 37.0 | 26-04-16 |
| 206373 | M1 Pro (16c) | 32 GB | gemma-4-26b-a4b-it | 4bit | 4k | 249.4 | 34.4 | 26-04-16 |
| 206374 | M4 Max (40c) | 48 GB | gemma-4-26b-a4b-it | 4bit | 1k | 1,160 | 99.8 | 26-04-16 |
| 206375 | M2 (10c) | 16 GB | Qwen3.5-9B | 4bit | 8k | 108.2 | 16.9 | 26-04-16 |
| 206376 | M2 (10c) | 16 GB | Qwen3.5-9B | 4bit | 4k | 109.0 | 17.6 | 26-04-16 |
| 206377 | M4 (10c) | 16 GB | Qwen3.5-2B-Distilled-OPUS... | 8bit | 8k | 997.9 | 44.6 | 26-04-16 |
| 206378 | M4 (10c) | 16 GB | Qwen3.5-2B-Distilled-OPUS... | 8bit | 4k | 994.2 | 45.3 | 26-04-16 |
| 206379 | M4 (10c) | 16 GB | Qwen3.5-2B-Distilled-OPUS... | 8bit | 1k | 876.3 | 45.5 | 26-04-16 |
| 206380 | M5 Max (40c) | 128 GB | Qwen2.5-Coder-14B-Instruc... | 4bit | 4k | 1,785 | 51.2 | 26-04-16 |