← Back to community benchmarks

Qwen3.5-9B

M4 (10c) · 16 GB · 4bit · 2026-03-11
Performance
16k
tokens
209.9
PP tok/s
15.5
TG tok/s
78038
TTFT (ms)
7.5
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 16 GB
GPU Cores 10
Software
oMLX v0.2.7
macOS macOS 26.3.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 222.3 19.6 6.5 GB view
4k 222.8 18.8 6.8 GB view
8k 219.3 17.6 7.0 GB view
16k 209.9 15.5 7.5 GB current
32k 186.1 12.6 8.5 GB view
64k 158.5 7.7 10.6 GB view
Batching Results
Batch Size TG tok/s Speedup
19.6 1.00×
38.6 1.97×
48.3 2.46×
48.9 2.49×