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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 1 | M3 Ultra (80c) | 512 GB | GLM-5.2 | 4bit | 4k | 209.6 | 18.4 | 26-07-25 |
| 2 | M3 Ultra (80c) | 512 GB | GLM-5.2 | 4bit | 1k | 192.5 | 19.3 | 26-07-25 |
| 3 | M1 Max (32c) | 64 GB | ThinkingCap-Qwen3.6-27B | 6bit | 64k | 67.6 | 10.6 | 26-07-25 |
| 4 | M1 Max (32c) | 64 GB | ThinkingCap-Qwen3.6-27B | 6bit | 128k | 51.4 | 8.6 | 26-07-25 |
| 5 | M1 Max (32c) | 64 GB | ThinkingCap-Qwen3.6-27B | 6bit | 16k | 88.6 | 13.0 | 26-07-25 |
| 6 | M1 Max (32c) | 64 GB | ThinkingCap-Qwen3.6-27B | 6bit | 4k | 92.5 | 13.7 | 26-07-25 |
| 7 | M4 Max (40c) | 128 GB | Qwen3.6-27B-AEON-Ultimate... | 8bit | 1k | 247.8 | 16.5 | 26-07-25 |
| 8 | M4 Max (40c) | 128 GB | Qwen3.6-27B-AEON-Ultimate... | 8bit | 4k | 251.4 | 16.2 | 26-07-25 |
| 9 | M4 (10c) | 16 GB | Qwen3.5-4B-OptiQ | 4bit | 4k | 399.4 | 27.1 | 26-07-25 |
| 10 | M4 (10c) | 16 GB | Qwen3.5-4B-OptiQ | 4bit | 8k | 389.6 | 26.2 | 26-07-25 |