← Back to community benchmarks

Qwen3.5-9B

M3 Max (30c) · 36 GB · 4bit · 2026-04-15

Performance

1k

tokens

453.3

PP tok/s

49.3

TG tok/s

2259

TTFT (ms)

5.8

Peak mem (GB)

Hardware

Chip M3 Max (30c)

Memory 36 GB

GPU Cores 30

Software

oMLX v0.3.5.dev1

macOS macOS 26.2

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	453.3	49.3	5.8 GB	current
4k	461.9	48.9	6.5 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	49.3	1.00×
2×	74.5	1.51×
4×	90.6	1.84×