← Back to community benchmarks

Qwen3.6-27B-oQ4-mtp

M4 Max (40c) · 128 GB · 4bit · 2026-05-24

Performance

16k

tokens

157.8

PP tok/s

15.2

TG tok/s

103805

TTFT (ms)

20.4

Peak mem (GB)

Hardware

Chip M4 Max (40c)

Memory 128 GB

GPU Cores 40

Software

oMLX v0.3.9

macOS macOS 26.5

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	184.7	24.7	16.5 GB	view
4k	198.8	21.7	17.9 GB	view
8k	185.8	21.3	18.9 GB	view
16k	157.8	15.2	20.4 GB	current
32k	137.8	14.3	23.4 GB	view