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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 156251 | M5 Pro (20c) | 48 GB | Nemotron-3-Nano-Omni-30B-... | 4bit | 4k | 1,668 | 97.1 | 26-05-01 |
| 156252 | M5 Pro (20c) | 48 GB | Nemotron-3-Nano-Omni-30B-... | 4bit | 1k | 933.4 | 99.8 | 26-05-01 |
| 156253 | M4 Max (32c) | 36 GB | Qwen3.5-4B | 4bit | 4k | 1,211 | 109.3 | 26-05-01 |
| 156254 | M4 Max (32c) | 36 GB | Qwen3.5-4B | 4bit | 1k | 1,064 | 113.5 | 26-05-01 |
| 156255 | M2 Ultra (60c) | 64 GB | Qwen3.6-35B-A3B | 4bit | 32k | 1,027 | 65.0 | 26-05-01 |
| 156256 | M2 Ultra (60c) | 64 GB | Qwen3.6-35B-A3B | 4bit | 1k | 904.5 | 84.7 | 26-05-01 |
| 156257 | M5 (10c) | 32 GB | Qwen3.6-35B-A3B-TurboQuan... | 4bit | 32k | 538.8 | 30.3 | 26-05-01 |
| 156258 | M5 (10c) | 32 GB | Qwen3.6-35B-A3B-TurboQuan... | 4bit | 16k | 646.8 | 37.0 | 26-05-01 |
| 156259 | M5 (10c) | 32 GB | Qwen3.6-35B-A3B-TurboQuan... | 4bit | 8k | 696.2 | 42.2 | 26-05-01 |
| 156260 | M5 (10c) | 32 GB | Qwen3.6-35B-A3B-TurboQuan... | 4bit | 1k | 549.7 | 47.5 | 26-05-01 |