Community Benchmarks
Download oMLX →Real performance numbers from oMLX users around the world. Submit yours from the app (v0.2.6+).
Reset filters| # | Chip | RAM | Model | Quant | Ctx | PP tok/s | TG tok/s | Date ↓ |
|---|---|---|---|---|---|---|---|---|
| 91 | M3 Ultra (80c) | 512 GB | Laguna-S-2.1-oQ5e | 5bit | 64k | 605.5 | 40.5 | 26-07-24 |
| 92 | M5 (10c) | 16 GB | gemma-4-e4b-it | 4bit | 64k | 1,339 | 22.3 | 26-07-24 |
| 93 | M1 Max (32c) | 64 GB | Laguna-XS-2.1 | 8bit | 64k | 266.9 | 27.3 | 26-07-24 |
| 94 | M3 Ultra (60c) | 256 GB | Qwen3.6-35B-A3B-oQ6-mtp | 6bit | 64k | 1,251 | 64.3 | 26-07-23 |
| 95 | M5 Max (40c) | 128 GB | Laguna-S-2.1-oQ6e | 6bit | 64k | 522.6 | 27.6 | 26-07-23 |
| 96 | M3 Max (40c) | 128 GB | gemma-4-12B-it | 4bit | 64k | 155.9 | 11.7 | 26-07-23 |
| 97 | M5 Max (40c) | 128 GB | Laguna-S-2.1-oQ4e | 4bit | 64k | 734.5 | 41.4 | 26-07-23 |
| 98 | M3 Max (40c) | 128 GB | Qwen3.5-9B | 4bit | 64k | 506.3 | 25.6 | 26-07-23 |
| 99 | M4 Pro (20c) | 48 GB | Qwen3.6-35B-A3B | 4bit | 64k | 524.0 | 48.4 | 26-07-23 |
| 100 | M1 Max (32c) | 64 GB | LFM2.5-8B-A1B | bf16 | 64k | 1,093 | 51.4 | 26-07-23 |