← Back to community benchmarks

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

M5 Max (40c) · 48 GB · 4bit · 2026-05-23

Performance

32k

tokens

698.4

PP tok/s

38.6

TG tok/s

46918

TTFT (ms)

22.8

Peak mem (GB)

Hardware

Chip M5 Max (40c)

Memory 48 GB

GPU Cores 40

Software

oMLX v0.3.9

macOS macOS 26.5

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	804.2	47.6	16.9 GB	view
4k	929.7	48.5	18.4 GB	view
8k	870.3	45.4	19.0 GB	view
16k	811.7	43.9	20.3 GB	view
32k	698.4	38.6	22.8 GB	current
64k	611.7	31.8	27.8 GB	view