← Back to community benchmarks

Qwen3.6-35B-A3B

M4 Pro (20c) · 64 GB · 4bit · 2026-04-16

Performance

16k

tokens

805.7

PP tok/s

66.1

TG tok/s

20335

TTFT (ms)

21.0

Peak mem (GB)

Hardware

Chip M4 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.3.5

macOS macOS 26.3

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	739.8	77.5	19.3 GB	view
4k	874.3	75.8	20.0 GB	view
8k	861.7	71.8	20.4 GB	view
16k	805.7	66.1	21.0 GB	current
32k	693.0	58.5	22.4 GB	view
64k	531.2	45.6	25.0 GB	view