llamaperf

Qwen3.6 35B (3B active)

on M3 Max 36GB · mlx_lm.server

Apple hardware
Nov 20, 2026
Throughput
41.5 t/s gen
Quant
4bit (MLX)
System RAM
36 GB

Summary

User benchmarks Qwen3.6-35B-A3B at 41.5 tok/s single-stream decode on an Apple M3 Max 36GB. Setup is mlx_lm.server with 4-bit MLX weights and continuous batching; prefill at 1k input is 919 tok/s. Decode peaks at 95 tok/s at 8 concurrent requests. Schema following scored 1.00, GSM8K 0.95 and IFEval 0.82 over 40 items.