← Back to community benchmarks

Qwen3.5-4B

M5 Pro (20c) · 48 GB · 4bit · 2026-03-21
Performance
8k
tokens
2,514
PP tok/s
80.9
TG tok/s
3259
TTFT (ms)
5.2
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 2,568 91.4 4.0 GB view
4k 2,564 85.7 4.8 GB view
8k 2,514 80.9 5.2 GB current
Batching Results
Batch Size TG tok/s Speedup
91.4 1.00×
161.2 1.76×
221.8 2.43×