← Back to community benchmarks

Qwen3.5-4B-oQ4-mtp

M4 Pro (16c) · 24 GB · 4bit · 2026-05-15

Performance

1k

tokens

592.7

PP tok/s

80.0

TG tok/s

1728

TTFT (ms)

3.6

Peak mem (GB)

Hardware

Chip M4 Pro (16c)

Memory 24 GB

GPU Cores 16

Software

oMLX v0.3.9.dev1

macOS macOS 26.4.1

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	592.7	80.0	3.6 GB	current
4k	635.1	76.7	4.3 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	80.0	1.00×
2×	120.0	1.50×
4×	129.2	1.61×