← Back to community benchmarks

Qwen3.5-9B

M4 (10c) · 24 GB · 4bit · 2026-03-10

Performance

8k

tokens

127.4

PP tok/s

10.8

TG tok/s

64327

TTFT (ms)

7.0

Peak mem (GB)

Hardware

Chip M4 (10c)

Memory 24 GB

GPU Cores 10

Software

oMLX v0.2.6

macOS macOS 26.3.1

Context 8,192

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	194.8	19.8	6.5 GB	view
4k	186.0	19.3	6.8 GB	view
8k	127.4	10.8	7.0 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	19.8	1.00×
2×	17.5	0.88×
4×	21.7	1.10×