← Back to community benchmarks

Qwen3-Coder-30B-A3B-Instruct

M5 Pro (20c) · 64 GB · 4bit · 2026-07-04

Performance

32k

tokens

1,069

PP tok/s

40.3

TG tok/s

30664

TTFT (ms)

19.6

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.4.4

macOS macOS 26.5.2

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,227	76.9	16.6 GB	view
8k	1,779	75.3	17.5 GB	view
32k	1,069	40.3	19.6 GB	current