← Back to community benchmarks

Qwen3.5-27B-heretic

M5 Pro (16c) · 48 GB · 8bit · 2026-04-18

Performance

4k

tokens

349.6

PP tok/s

9.7

TG tok/s

11715

TTFT (ms)

29.8

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 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	317.2	9.8	28.3 GB	view
4k	349.6	9.7	29.8 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	9.8	1.00×
2×	17.3	1.77×
4×	22.2	2.27×