Qwen3-Coder 30B (3B active)
on M1 Max 64GB · llama.cpp
Oct 12, 2026
Summary
User reports Qwen3-Coder 30B-A3B at 64.4 t/s generation and 645 t/s prompt processing on an M1 Max 64GB.
Setup is llama.cpp with Q4_K_M, measured via llama-bench on Metal.
The figure is a native prefill/decode rate, not a TTFT-derived estimate; exact test lengths, context, batch settings and KV precision are not given.