← Back to community benchmarks

Qwen3.5-4B

M4 (10c) · 16 GB · 4bit · 2026-04-15
Performance
8k
tokens
390.1
PP tok/s
34.8
TG tok/s
20997
TTFT (ms)
4.3
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 16 GB
GPU Cores 10
Software
oMLX v0.3.5.dev1
macOS macOS 26.2
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 382.9 39.0 3.3 GB view
4k 398.0 37.5 3.9 GB view
8k 390.1 34.8 4.3 GB current
Batching Results
Batch Size TG tok/s Speedup
39.0 1.00×
72.8 1.87×
118.3 3.03×
289.4 7.42×