← Back to community benchmarks

Qwen3.5-9B

M5 Pro (20c) · 48 GB · 4bit · 2026-03-19
Performance
8k
tokens
1,608
PP tok/s
52.3
TG tok/s
5093
TTFT (ms)
7.8
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 48 GB
GPU Cores 20
Software
oMLX v0.2.19
macOS macOS 26.3.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,532 54.7 6.7 GB view
4k 1,615 53.9 7.4 GB view
8k 1,608 52.3 7.8 GB current
Batching Results
Batch Size TG tok/s Speedup
54.7 1.00×
102.0 1.86×
137.5 2.51×
147.7 2.70×