← Back to community benchmarks

Qwen3.5-4B

M4 (10c) · 32 GB · 4bit · 2026-03-11
Performance
4k
tokens
351.6
PP tok/s
29.1
TG tok/s
11649
TTFT (ms)
4.2
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.2.7
macOS macOS 26.3.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 372.1 30.8 3.9 GB view
4k 351.6 29.1 4.2 GB current
Batching Results
Batch Size TG tok/s Speedup
30.8 1.00×
57.3 1.86×
73.8 2.40×