← Back to community benchmarks

Qwen3.5-9B

M4 (10c) · 32 GB · 4bit · 2026-03-10
Performance
8k
tokens
215.6
PP tok/s
19.3
TG tok/s
37997
TTFT (ms)
7.0
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.2.7
macOS macOS 15.7.4
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 224.1 21.6 6.5 GB view
8k 215.6 19.3 7.0 GB current
16k 210.3 17.3 7.5 GB view
32k 190.7 13.9 8.5 GB view
64k 164.0 9.9 10.6 GB view
Batching Results
Batch Size TG tok/s Speedup
21.6 1.00×
41.9 1.94×
50.5 2.34×
52.2 2.42×