← Back to community benchmarks

Qwen3.6-35B-A3B8bit

M2 Max (38c) · 96 GB · 8bit · 2026-05-04

Performance

16k

tokens

1,088

PP tok/s

109.9

TG tok/s

15065

TTFT (ms)

43.1

Peak mem (GB)

Hardware

Chip M2 Max (38c)

Memory 96 GB

GPU Cores 38

Software

oMLX v0.3.8

macOS macOS 26.4.1

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,179	203.6	36.8 GB	view
4k	1,177	175.8	40.5 GB	view
8k	1,167	146.8	41.4 GB	view
16k	1,088	109.9	43.1 GB	current
32k	941.5	75.2	46.4 GB	view
64k	718.0	45.8	52.8 GB	view
128k	472.5	23.5	65.7 GB	view
195k	306.4	15.5	78.9 GB	view