← Back to community benchmarks

Qwen36-27B-oQ6-mtp

M4 Max (40c) · 128 GB · 6bit · 2026-05-20

Performance

32k

tokens

221.7

PP tok/s

24.2

TG tok/s

147792

TTFT (ms)

29.7

Peak mem (GB)

Hardware

Chip M4 Max (40c)

Memory 128 GB

GPU Cores 40

Software

oMLX v0.3.9rc1

macOS macOS 15.7.4

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	234.6	27.7	22.8 GB	view
4k	243.1	28.0	24.2 GB	view
8k	241.4	27.5	25.2 GB	view
16k	235.1	26.3	26.7 GB	view
32k	221.7	24.2	29.7 GB	current