← Back to community benchmarks

Qwen3.5-35B-A3B

M3 Pro (18c) · 36 GB · 4bit · 2026-04-13

Performance

4k

tokens

648.5

PP tok/s

50.9

TG tok/s

6316

TTFT (ms)

20.0

Peak mem (GB)

Hardware

Chip M3 Pro (18c)

Memory 36 GB

GPU Cores 18

Software

oMLX v0.3.5.dev1

macOS macOS 26.4.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	569.7	52.8	19.3 GB	view
4k	648.5	50.9	20.0 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	52.8	1.00×
2×	79.0	1.50×
4×	104.1	1.97×
8×	131.6	2.49×