← Back to community benchmarks

Qwen3.5-27B-Heretic

M4 Max (40c) · 128 GB · 4bit · 2026-03-10

Performance

4k

tokens

203.9

PP tok/s

28.0

TG tok/s

20090

TTFT (ms)

18.3

Peak mem (GB)

Hardware

Chip M4 Max (40c)

Memory 128 GB

GPU Cores 40

Software

oMLX v0.2.7

macOS macOS 26.3.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	231.2	29.8	17.5 GB	view
4k	203.9	28.0	18.3 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	29.8	1.00×
2×	50.6	1.70×
4×	53.3	1.79×