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 ↓ |
|---|---|---|---|---|---|---|---|---|
| 75121 | M4 (10c) | 16 GB | Qwen3.5-0.8B | 4bit | 4k | 2,147 | 166.4 | 26-03-09 |
| 75122 | M4 (10c) | 16 GB | Qwen3.5-0.8B | 4bit | 1k | 2,150 | 178.0 | 26-03-09 |
| 75123 | M1 Ultra (48c) | 128 GB | Huihui-Qwen3.5-35B-A3B-ab... | 8bit | 4k | 851.4 | 51.6 | 26-03-09 |
| 75124 | M4 (10c) | 16 GB | Qwen3.5-9B | 4bit | 1k | 194.2 | 16.8 | 26-03-09 |
| 75125 | M4 (10c) | 16 GB | Qwen3.5-9B | 4bit | 4k | 191.8 | 20.1 | 26-03-09 |
| 75126 | M2 Max (38c) | 64 GB | Qwen3.5-35B-A3B | 4bit | 4k | 742.3 | 75.8 | 26-03-09 |
| 75127 | M2 Max (38c) | 64 GB | Qwen3.5-35B-A3B | 4bit | 1k | 681.9 | 78.8 | 26-03-09 |
| 75128 | M1 Ultra (48c) | 128 GB | Huihui-Qwen3.5-35B-A3B-ab... | 8bit | 1k | 821.3 | 53.2 | 26-03-09 |
| 75129 | M1 Pro (16c) | 16 GB | Qwen3-VL-8B-Instruct | 4bit | 8k | 145.2 | 22.6 | 26-03-09 |
| 75130 | M1 Pro (16c) | 16 GB | Qwen3-VL-8B-Instruct | 4bit | 4k | 153.8 | 27.1 | 26-03-09 |