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 ↓
5801 M3 Ultra (60c) 96 GB Qwen3.5-122B-A10B 4bit 16k 670.7 35.9 26-03-08
5802 M3 Ultra (60c) 96 GB Qwen3.5-122B-A10B 4bit 32k 587.5 24.4 26-03-08
5803 M3 Ultra (60c) 96 GB Qwen3.5-122B-A10B 4bit 8k 712.7 47.3 26-03-08
5804 M3 Ultra (60c) 96 GB Qwen3.5-122B-A10B 4bit 4k 714.1 50.9 26-03-08
5805 M3 Ultra (60c) 96 GB Qwen3.5-122B-A10B 4bit 1k 638.7 54.4 26-03-08
5806 M4 (10c) 24 GB Qwen3.5-9B-mxfp4 4bit 4k 169.0 20.7 26-03-08
5807 M4 (10c) 24 GB Qwen3.5-9B-mxfp4 4bit 1k 169.7 21.7 26-03-08
5808 M2 Ultra (76c) 128 GB Qwen3.5-27B 4bit 64k 192.2 16.8 26-03-08
5809 M2 Ultra (76c) 128 GB Qwen3.5-27B 4bit 32k 212.9 21.9 26-03-08
5810 M2 Ultra (76c) 128 GB Qwen3.5-27B 4bit 8k 227.0 29.3 26-03-08
5,957 results · Page 581 of 596