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 ↓
351 M3 Ultra (80c) 512 GB Qwen3-Coder-Next 8bit 4k 1,801 58.6 26-03-10
352 M4 (10c) 32 GB Qwen3.5-9B 4bit 1k 224.9 21.8 26-03-10
353 M3 Ultra (80c) 512 GB Qwen3-Coder-Next 8bit 8k 1,884 55.9 26-03-10
354 M3 Ultra (80c) 512 GB Qwen3-Coder-Next 8bit 1k 1,245 61.1 26-03-10
355 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 4bit 32k 313.4 21.7 26-03-10
356 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 4bit 64k 237.6 13.3 26-03-10
357 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 4bit 16k 344.8 30.8 26-03-10
358 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 4bit 4k 422.0 44.3 26-03-10
359 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 4bit 8k 356.5 42.7 26-03-10
360 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 4bit 1k 554.6 52.6 26-03-10
4,380 results · Page 36 of 438