← Back to community benchmarks

Qwen3.6-27B-PARO

M4 Max (40c) · 64 GB · 4bit · 2026-05-19

Performance

1k

tokens

249.8

PP tok/s

28.7

TG tok/s

4100

TTFT (ms)

15.5

Peak mem (GB)

Hardware

Chip M4 Max (40c)

Memory 64 GB

GPU Cores 40

Software

oMLX v0.3.9rc1

macOS macOS 26.5

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	249.8	28.7	15.5 GB	current
4k	241.5	27.6	16.7 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	28.7	1.00×
2×	49.7	1.73×
4×	58.1	2.02×