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 ↓
330071 M2 Max (30c) 32 GB Qwen3.5-0.8B 8bit 4k 3,365 211.6 26-03-10
330072 M2 Max (30c) 32 GB Qwen3.5-0.8B 8bit 1k 3,360 235.0 26-03-10
330073 M2 Max (30c) 32 GB Qwen3.5-0.8B 8bit 4k 3,077 211.0 26-03-10
330074 M2 Max (30c) 32 GB Qwen3.5-0.8B 8bit 1k 2,806 238.8 26-03-10
330075 M3 Pro (18c) 36 GB Qwen3-4B-Instruct-25075bi... 5bit 4k 529.7 33.5 26-03-10
330076 M3 Pro (18c) 36 GB Qwen3-4B-Instruct-25075bi... 5bit 1k 575.7 38.0 26-03-10
330077 M4 Pro (20c) 24 GB Qwen3.5-4B 4bit 4k 616.1 44.8 26-03-10
330078 M4 Pro (20c) 24 GB Qwen3.5-4B 4bit 1k 595.3 28.1 26-03-10
330079 M4 (10c) 16 GB Qwen3.5-9B 4bit 4k 212.9 18.4 26-03-10
330080 M4 (10c) 16 GB Qwen3.5-9B 4bit 1k 209.8 17.9 26-03-10
334,259 results · Page 33008 of 33426