← Back to community benchmarks

Qwen3.6-35B-A3B

M3 Ultra (80c) · 512 GB · 8bit · 2026-04-23

Performance

16k

tokens

2,502

PP tok/s

66.3

TG tok/s

6549

TTFT (ms)

37.1

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 512 GB

GPU Cores 80

Software

oMLX v0.3.7

macOS macOS 26.3

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,602	294.1	36.7 GB	view
4k	713.4	74.7	36.2 GB	view
8k	2,641	72.5	36.5 GB	view
16k	2,502	66.3	37.1 GB	current