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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 71021 | M3 Max (40c) | 48 GB | Qwopus3.5-9B-v3 | 4bit | 1k | 654.9 | 69.6 | 26-05-20 |
| 71022 | M5 (10c) | 32 GB | Qwen3.6-27B-oQ4-mtp | 4bit | 16k | 143.3 | 9.0 | 26-05-20 |
| 71023 | M5 (10c) | 32 GB | Qwen3.6-27B-oQ4-mtp | 4bit | 8k | 178.5 | 11.9 | 26-05-20 |
| 71024 | M5 (10c) | 32 GB | Qwen3.6-27B-oQ4-mtp | 4bit | 4k | 194.6 | 13.2 | 26-05-20 |
| 71025 | M5 (10c) | 32 GB | Qwen3.6-27B-oQ4-mtp | 4bit | 1k | 190.3 | 13.4 | 26-05-20 |
| 71026 | M5 (10c) | 24 GB | gemma-4-e4b-it | 4bit | 4k | 1,706 | 26.3 | 26-05-20 |
| 71027 | M5 (10c) | 24 GB | gemma-4-e4b-it | 4bit | 1k | 1,304 | 27.3 | 26-05-20 |
| 71028 | M1 Ultra (48c) | 64 GB | Qwen3.6-35B-A3B | 4bit | 4k | 867.7 | 62.6 | 26-05-20 |
| 71029 | M1 Ultra (48c) | 64 GB | Qwen3.6-35B-A3B | 4bit | 1k | 585.0 | 65.7 | 26-05-20 |
| 71030 | M4 Max (40c) | 48 GB | Qwen3.6-27B-oQ8-mtp | 8bit | 1k | 237.2 | 26.9 | 26-05-20 |