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 ↓
155751 M5 Max (40c) 128 GB Qwen3.6-27B-UD 4bit 1k 762.9 21.3 26-04-29
155752 M1 Pro (16c) 32 GB Qwen3.6-35B-A3B-UD 4bit 4k 306.2 37.6 26-04-29
155753 M1 Pro (16c) 32 GB Qwen3.6-35B-A3B-UD 4bit 1k 266.8 34.5 26-04-29
155754 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 64k 1,412 38.2 26-04-29
155755 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 32k 1,711 49.9 26-04-29
155756 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 16k 1,845 58.4 26-04-29
155757 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 8k 1,855 61.7 26-04-29
155758 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 4k 1,793 62.3 26-04-29
155759 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 8bit 1k 1,319 65.4 26-04-29
155760 M4 Max (40c) 128 GB Qwen3.6-35B-A3B-UD 4bit 32k 1,206 52.4 26-04-29
310,048 results · Page 15576 of 31005