← Back to community benchmarks

Qwen3.5-35B-A3B

M4 Max (40c) · 48 GB · 4bit · 2026-03-21

Performance

16k

tokens

1,581

PP tok/s

86.1

TG tok/s

10365

TTFT (ms)

22.0

Peak mem (GB)

Hardware

Chip M4 Max (40c)

Memory 48 GB

GPU Cores 40

Software

oMLX v0.2.17

macOS macOS 26.3.1

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
4k	1,809	101.5	21.0 GB	view
8k	1,755	96.0	21.4 GB	view
16k	1,581	86.1	22.0 GB	current
32k	1,343	70.6	23.4 GB	view
64k	988.7	48.5	26.0 GB	view
128k	676.8	30.1	31.4 GB	view
195k	436.9	5.6	37.0 GB	view