← Back to community benchmarks

Qwen3.6-35B-A3B-oQ4

M5 Max (40c) · 128 GB · 4bit · 2026-04-20

Performance

8k

tokens

3,885

PP tok/s

104.7

TG tok/s

2109

TTFT (ms)

21.1

Peak mem (GB)

Hardware

Chip M5 Max (40c)

Memory 128 GB

GPU Cores 40

Software

oMLX v0.3.6

macOS macOS 26.4.1

Context 8,192

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	2,734	436.7	21.3 GB	view
4k	1,331	112.8	20.7 GB	view
8k	3,885	104.7	21.1 GB	current
16k	3,621	95.9	21.7 GB	view
32k	2,642	73.9	23.0 GB	view
64k	1,975	47.6	25.7 GB	view