← Back to community benchmarks

Qwen3.5-9B-OptiQ

M5 Pro (20c) · 24 GB · 4bit · 2026-05-13
Performance
32k
tokens
1,251
PP tok/s
36.4
TG tok/s
26188
TTFT (ms)
11.6
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 24 GB
GPU Cores 20
Software
oMLX v0.3.9.dev1
macOS macOS 26.5
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,204 43.4 7.7 GB view
4k 1,539 42.8 8.4 GB view
8k 1,570 41.6 9.0 GB view
16k 1,514 40.3 9.8 GB view
32k 1,251 36.4 11.6 GB current
Batching Results
Batch Size TG tok/s Speedup
43.4 1.00×
76.9 1.77×
108.9 2.51×