← Back to community benchmarks

Qwen3.6-35B-A3B-MTP

M3 Ultra (80c) · 96 GB · 4bit · 2026-07-12

Performance

32k

tokens

2,118

PP tok/s

83.0

TG tok/s

15470

TTFT (ms)

20.5

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 96 GB

GPU Cores 80

Software

oMLX v0.5.0

macOS macOS 26.5.2

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
32k	2,118	83.0	20.5 GB	current
64k	1,613	72.8	21.6 GB	view
128k	1,090	61.7	22.6 GB	view