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 ↓
206231 M5 Pro (20c) 64 GB Qwen3-30B-A3B-Instruct-25... 4bit 4k 2,496 75.3 26-04-16
206232 M5 (8c) 16 GB Qwen3.5-4B 4bit 1k 913.7 47.8 26-04-16
206233 M5 (8c) 16 GB Qwen3.5-4B 4bit 4k 1,004 45.8 26-04-16
206234 M5 Pro (20c) 64 GB Qwen3-30B-A3B-Instruct-25... 4bit 4k 2,492 75.3 26-04-16
206235 M5 Pro (20c) 64 GB Qwen3-30B-A3B-Instruct-25... 4bit 1k 1,862 92.1 26-04-16
206236 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 4bit 1k 998.7 107.1 26-04-16
206237 M5 Pro (20c) 64 GB NVIDIA-Nemotron-3-Nano-30... 4bit 4k 1,832 103.6 26-04-16
206238 M1 Max (32c) 64 GB Qwen3.5-35B-A3B 4bit 4k 383.5 60.8 26-04-16
206239 M1 Max (32c) 64 GB Qwen3.5-35B-A3B 4bit 1k 378.8 113.3 26-04-16
206240 M3 Max (40c) 128 GB Qwen3.5-122B-A10B-oQ4 4bit 1k 439.0 40.0 26-04-16
304,856 results · Page 20624 of 30486