← Back to community benchmarks

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

M5 Pro (20c) · 64 GB · 4bit · 2026-07-21

Performance

64k

tokens

628.6

PP tok/s

50.1

TG tok/s

104258

TTFT (ms)

22.4

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.5.2.dev1

macOS macOS 26.5.2

Context 65,536

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	2,046	71.7	20.5 GB	view
4k	2,446	71.8	21.4 GB	view
8k	2,374	69.7	21.7 GB	view
16k	1,510	66.3	21.5 GB	view
32k	1,020	59.8	21.8 GB	view
64k	628.6	50.1	22.4 GB	current