← Back to community benchmarks

Qwen3.5-4B-OptiQ

M4 (10c) · 16 GB · 4bit · 2026-03-10
Performance
8k
tokens
339.6
PP tok/s
20.9
TG tok/s
24123
TTFT (ms)
5.7
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 16 GB
GPU Cores 10
Software
oMLX v0.2.6
macOS macOS 26.3.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 334.9 24.0 4.8 GB view
4k 349.1 22.6 5.4 GB view
8k 339.6 20.9 5.7 GB current
16k 315.0 16.9 6.7 GB view
Batching Results
Batch Size TG tok/s Speedup
24.0 1.00×
45.5 1.90×
70.5 2.94×
139.8 5.83×