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 ↓
371 M3 Ultra (60c) 256 GB Qwen3-Coder-Next 4bit 4k 1,556 70.4 26-03-10
372 M3 Ultra (60c) 256 GB Qwen3-Coder-Next 4bit 8k 1,553 67.5 26-03-10
373 M3 Ultra (60c) 256 GB Qwen3-Coder-Next 4bit 1k 1,318 73.0 26-03-10
374 M1 (8c) 16 GB Qwen2.5-7B-Instruct 4bit 4k 136.3 13.3 26-03-10
375 M1 (8c) 16 GB Qwen2.5-7B-Instruct 4bit 1k 140.0 14.4 26-03-10
376 M4 (10c) 16 GB Qwen3.5-9B 4bit 1k 201.0 18.2 26-03-10
377 M1 Pro (14c) 32 GB Qwen3.5-35B-A3Blm 4bit 4k 264.3 41.4 26-03-10
378 M2 Ultra (60c) 192 GB Qwen3.5-122B-A10B 8bit 1k 347.3 37.0 26-03-10
379 M2 Ultra (60c) 192 GB Qwen3.5-122B-A10B 8bit 1k 346.8 37.1 26-03-10
380 M3 Max (40c) 128 GB Qwen3.5-27B 4bit 4k 189.9 19.2 26-03-10
4,970 results · Page 38 of 497