← Back to community benchmarks

Qwen3.5-9B-MTPLX-Optimized-Speed

M5 Pro (20c) · 64 GB · 6bit · 2026-07-24

Performance

16k

tokens

1,423

PP tok/s

34.7

TG tok/s

11513

TTFT (ms)

9.7

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.5.3

macOS macOS 26.5.2

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
4k	1,522	36.8	8.6 GB	view
16k	1,423	34.7	9.7 GB	current
32k	1,241	32.3	11.2 GB	view
64k	1,008	27.7	14.2 GB	view
128k	678.5	20.5	20.6 GB	view
195k	495.2	16.4	27.2 GB	view