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 ↓
65571 M1 Pro (14c) 16 GB Qwen3.5-9B 4bit 1k 137.5 29.9 26-03-10
65572 M2 Max (38c) 96 GB Llama-3.3-70B-Instruct 4bit 4k 64.7 7.7 26-03-10
65573 M2 Max (38c) 96 GB Llama-3.3-70B-Instruct 4bit 1k 72.7 8.6 26-03-10
65574 M4 Max (40c) 128 GB Huihui-Qwen3-Next-80B-A3B... 4bit 4k 1,300 85.9 26-03-10
65575 M4 Max (40c) 128 GB Huihui-Qwen3-Next-80B-A3B... 4bit 1k 1,044 89.2 26-03-10
65576 M4 (10c) 16 GB Qwen3.5-4B 4bit 4k 399.5 32.5 26-03-10
65577 M4 (10c) 16 GB Qwen3.5-4B 4bit 1k 402.8 34.5 26-03-10
65578 M3 Ultra (60c) 96 GB Qwen3.5-4B unknown 1k 986.8 17.3 26-03-10
65579 M4 (10c) 16 GB Qwen3.5-4B 4bit 1k 396.8 34.3 26-03-10
65580 M4 (10c) 32 GB Qwen3.5-4B 4bit 4k 404.6 35.0 26-03-10
67,646 results · Page 6558 of 6765