← Back to community benchmarks

Qwen3.5-4B

M3 Ultra (80c) · 512 GB · 4bit · 2026-04-30

Performance

4k

tokens

2,350

PP tok/s

152.2

TG tok/s

1743

TTFT (ms)

4.1

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 512 GB

GPU Cores 80

Software

oMLX v0.3.6

macOS macOS 26.4.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,681	161.6	3.4 GB	view
4k	2,350	152.2	4.1 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	161.6	1.00×
2×	273.3	1.69×
4×	373.0	2.31×