Community Benchmarks

Download oMLX →

Real performance numbers from oMLX users around the world. Submit yours from the app (v0.2.6+).

Reset filters
# Chip RAM Model Quant Ctx PP tok/s TG tok/s Date ↓
2091 M2 Ultra (60c) 192 GB Qwen3.5-122B-A10B 4bit 1k 351.9 44.4 26-03-20
2092 M3 Ultra (60c) 256 GB Qwen3.5-122B-A10B 8bit 8k 727.3 37.9 26-03-20
2093 M3 Ultra (60c) 256 GB Qwen3.5-122B-A10B 8bit 4k 740.9 39.1 26-03-20
2094 M3 Ultra (60c) 256 GB Qwen3.5-122B-A10B 8bit 1k 676.1 41.9 26-03-20
2095 M3 Ultra (80c) 512 GB Qwen3.5-122B-A10B 8bit 4k 906.8 38.4 26-03-20
2096 M3 Ultra (80c) 512 GB Qwen3.5-122B-A10B 8bit 1k 836.1 39.6 26-03-20
2097 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 6bit 4k 548.6 40.9 26-03-19
2098 M4 Max (40c) 128 GB Qwen3.5-122B-A10B 6bit 1k 505.1 40.2 26-03-19
2099 M4 Max (40c) 128 GB Qwen3.5-122B-A10B-5bit 5bit 1k 543.6 48.9 26-03-19
2100 M4 Max (40c) 128 GB Qwen3.5-122B-A10B-5bit 5bit 4k 578.7 46.8 26-03-19
2,971 results · Page 210 of 298