← Back to community benchmarks

Qwen3.6-27B

M5 Pro (16c) · 48 GB · 4bit · 2026-04-26

Performance

16k

tokens

347.9

PP tok/s

15.5

TG tok/s

47095

TTFT (ms)

19.1

Peak mem (GB)

Hardware

Chip M5 Pro (16c)

Memory 48 GB

GPU Cores 16

Software

oMLX v0.3.6

macOS macOS 26.4.1

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	347.8	17.1	15.8 GB	view
4k	378.5	17.1	17.3 GB	view
16k	347.9	15.5	19.1 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	17.1	1.00×
2×	26.5	1.55×
4×	31.9	1.87×