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 ↓
155071 M3 Max (40c) 48 GB Qwen3.6-35B-A3B-heretic-o... 6bit 4k 358.3 69.9 26-04-27
155072 M3 Max (40c) 48 GB Qwen3.6-35B-A3B-heretic-o... 6bit 1k 1,098 214.2 26-04-27
155073 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 64k 1,498 63.8 26-04-27
155074 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 32k 1,846 66.0 26-04-27
155075 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 16k 1,989 66.7 26-04-27
155076 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 8k 1,910 64.3 26-04-27
155077 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 4k 1,864 66.5 26-04-27
155078 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 1k 1,363 67.7 26-04-27
155079 M3 Max (30c) 36 GB Qwen3.6-35B-A3B 4bit 4k 1,105 81.3 26-04-27
155080 M3 Max (30c) 36 GB Qwen3.6-35B-A3B 4bit 1k 779.3 80.4 26-04-27
301,851 results · Page 15508 of 30186