← Back to community benchmarks

Qwen3-4B-Thinking-2507

M2 (10c) · 8 GB · 4bit · 2026-03-14

Performance

1k

tokens

193.8

PP tok/s

31.9

TG tok/s

5285

TTFT (ms)

2.8

Peak mem (GB)

Hardware

Chip M2 (10c)

Memory 8 GB

GPU Cores 10

Software

oMLX v0.2.11

macOS macOS 26.2

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	193.8	31.9	2.8 GB	current
4k	171.7	25.4	3.2 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	31.9	1.00×
2×	42.2	1.32×
4×	32.5	1.02×