← Back to community benchmarks

Qwen3.5-4B

M5 Max (40c) · 128 GB · 4bit · 2026-04-02
Performance
1k
tokens
2,819
PP tok/s
165.3
TG tok/s
363
TTFT (ms)
3.4
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.0
macOS macOS 26.4
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,819 165.3 3.4 GB current
4k 4,336 153.0 4.2 GB view
8k 4,644 141.9 4.6 GB view
16k 4,535 123.3 5.3 GB view
Batching Results
Batch Size TG tok/s Speedup
165.3 1.00×
292.2 1.77×
412.6 2.50×
473.8 2.87×