← Back to community benchmarks

Qwen3-Coder-30B-A3B-Instruct

M5 Pro (20c) · 64 GB · 8bit · 2026-03-16

Performance

1k

tokens

1,505

PP tok/s

58.3

TG tok/s

680

TTFT (ms)

30.8

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.2.13

macOS macOS 26.3

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,505	58.3	30.8 GB	current
4k	1,698	52.5	31.3 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	58.3	1.00×
2×	84.4	1.45×
4×	104.6	1.79×