← Back to community benchmarks

Qwen3.6-27B-oQ6-mtp

M4 Pro (20c) · 64 GB · 6bit · 2026-05-13

Performance

1k

tokens

81.7

PP tok/s

6.7

TG tok/s

12541

TTFT (ms)

22.6

Peak mem (GB)

Hardware

Chip M4 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.3.9.dev2

macOS macOS 26.4.1

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	81.7	6.7	22.6 GB	current
4k	79.1	6.5	24.0 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	6.7	1.00×
2×	9.4	1.40×
4×	11.4	1.70×