← Back to community benchmarks

Qwen3.6-27B4bit-MTP

M2 Ultra (60c) · 64 GB · 4bit · 2026-05-20

Performance

8k

tokens

179.6

PP tok/s

30.8

TG tok/s

45604

TTFT (ms)

18.5

Peak mem (GB)

Hardware

Chip M2 Ultra (60c)

Memory 64 GB

GPU Cores 60

Software

oMLX v0.3.9rc1

macOS macOS 26.3.1

Context 8,192

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	172.3	29.2	16.1 GB	view
4k	180.1	31.1	17.5 GB	view
8k	179.6	30.8	18.5 GB	current