← Back to community benchmarks

Qwen3.5-27B-oQ4

M5 Pro (20c) · 64 GB · 4bit · 2026-03-25

Performance

1k

tokens

327.0

PP tok/s

12.7

TG tok/s

3131

TTFT (ms)

18.8

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.2.20

macOS macOS 26.3.1

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	327.0	12.7	18.8 GB	current
4k	330.6	12.3	21.9 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	12.7	1.00×
2×	23.3	1.83×
4×	34.3	2.70×