← Back to community benchmarks

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

M1 Max (32c) · 32 GB · 4bit · 2026-05-14

Performance

4k

tokens

150.5

PP tok/s

22.7

TG tok/s

27225

TTFT (ms)

18.2

Peak mem (GB)

Hardware

Chip M1 Max (32c)

Memory 32 GB

GPU Cores 32

Software

oMLX v0.3.9.dev2

macOS macOS 26.2

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	146.9	23.0	16.7 GB	view
4k	150.5	22.7	18.2 GB	current
8k	142.8	18.9	19.2 GB	view