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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 206071 | M4 (10c) | 16 GB | Qwen3.5-9B | 4bit | 4k | 222.7 | 20.4 | 26-04-15 |
| 206072 | M5 Max (40c) | 128 GB | Qwen3.5-122B-A10B | 4bit | 8k | 800.4 | 46.5 | 26-04-15 |
| 206073 | M5 Max (40c) | 128 GB | Qwen3.5-122B-A10B | 4bit | 16k | 851.3 | 47.6 | 26-04-15 |
| 206074 | M5 Max (40c) | 128 GB | Qwen3.5-122B-A10B | 4bit | 1k | 426.4 | 44.2 | 26-04-15 |
| 206075 | M5 Max (40c) | 128 GB | Qwen3.5-122B-A10B | 4bit | 4k | 718.7 | 46.5 | 26-04-15 |
| 206076 | M1 Ultra (48c) | 128 GB | gemma-4-26b-a4b-it | 4bit | 1k | 601.9 | 60.6 | 26-04-15 |
| 206077 | M1 Ultra (48c) | 128 GB | gemma-4-26b-a4b-it | 4bit | 4k | 718.3 | 59.0 | 26-04-15 |
| 206078 | M3 Max (30c) | 36 GB | Qwen3.5-9B | 4bit | 4k | 461.9 | 48.9 | 26-04-15 |
| 206079 | M3 Max (30c) | 36 GB | Qwen3.5-9B | 4bit | 1k | 453.3 | 49.3 | 26-04-15 |
| 206080 | M5 Max (40c) | 128 GB | Qwen3.5-27B-heretic-oQ4 | 4bit | 64k | 543.5 | 18.7 | 26-04-15 |