← Back to community benchmarks

Qwen3.5-4B

M2 (10c) · 8 GB · 4bit · 2026-03-10
Performance
1k
tokens
196.0
PP tok/s
30.3
TG tok/s
5223
TTFT (ms)
3.9
Peak mem (GB)
Hardware
Chip M2 (10c)
Memory 8 GB
GPU Cores 10
Software
oMLX v0.2.6
macOS macOS 26.2
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 196.0 30.3 3.9 GB current
4k 199.5 28.9 4.2 GB view
Batching Results
Batch Size TG tok/s Speedup
30.3 1.00×
44.5 1.47×
46.2 1.52×