← Back to community benchmarks

Qwen2.5-Coder-3B-Instruct

M3 Max (30c) · 36 GB · 4bit · 2026-03-11

Performance

1k

tokens

1,078

PP tok/s

96.9

TG tok/s

950

TTFT (ms)

2.4

Peak mem (GB)

Hardware

Chip M3 Max (30c)

Memory 36 GB

GPU Cores 30

Software

oMLX v0.2.7

macOS macOS 26.2

Context 1,024

Batching Results

Batch Size	TG tok/s	Speedup
1×	96.9	1.00×
2×	177.5	1.83×
4×	232.4	2.40×