← Back to community benchmarks

Qwen3.5-9B-Fable-5-v1-oQ8

M5 Pro (20c) · 48 GB · 8bit · 2026-07-20

Performance

64k

tokens

614.3

PP tok/s

22.8

TG tok/s

106677

TTFT (ms)

13.4

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 48 GB

GPU Cores 20

Software

oMLX v0.5.1

macOS macOS 26.4.1

Context 65,536

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
4k	1,547	29.7	10.6 GB	view
16k	1,127	28.5	11.1 GB	view
64k	614.3	22.8	13.4 GB	current