← Back to community benchmarks

unsloth--Qwen3.6-35B-A3B-UD

M5 Pro (20c) · 64 GB · 4bit · 2026-07-23

Performance

8k

tokens

2,365

PP tok/s

69.9

TG tok/s

3464

TTFT (ms)

21.7

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.5.3

macOS macOS 26.5.2

Context 8,192

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	2,039	74.2	20.5 GB	view
4k	2,434	71.1	21.4 GB	view
8k	2,365	69.9	21.7 GB	current
16k	2,129	66.5	22.4 GB	view
32k	1,763	60.3	23.7 GB	view
64k	1,267	50.2	26.3 GB	view