← Back to community benchmarks

Qwen3.6-27B-oQ4-fp16-mtp

M5 Max (32c) · 36 GB · 4bit · 2026-06-18

Performance

32k

tokens

503.8

PP tok/s

21.1

TG tok/s

65038

TTFT (ms)

20.9

Peak mem (GB)

Hardware

Chip M5 Max (32c)

Memory 36 GB

GPU Cores 32

Software

oMLX v0.4.4

macOS macOS 26.6

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
8k	661.5	23.0	18.5 GB	view
16k	561.7	20.4	19.4 GB	view
32k	503.8	21.1	20.9 GB	current