← Back to community benchmarks

Qwen3.5-9B

M4 (10c) · 16 GB · 4bit · 2026-03-11
Performance
4k
tokens
223.3
PP tok/s
21.3
TG tok/s
18343
TTFT (ms)
5.9
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 16 GB
GPU Cores 10
Software
oMLX v0.2.8
macOS macOS 26.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 215.4 21.8 5.7 GB view
4k 223.3 21.3 5.9 GB current
8k 220.8 20.2 6.2 GB view
16k 204.0 18.3 6.7 GB view
32k 187.8 15.6 7.7 GB view
Batching Results
Batch Size TG tok/s Speedup
21.8 1.00×
42.3 1.94×
50.9 2.33×