← Back to community benchmarks

Qwen3.5-9B

M5 (10c) · 24 GB · 4bit · 2026-03-10
Performance
16k
tokens
224.4
PP tok/s
21.6
TG tok/s
73019
TTFT (ms)
7.5
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 24 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 233.2 24.2 6.5 GB view
4k 236.1 24.4 6.8 GB view
8k 229.7 23.5 7.0 GB view
16k 224.4 21.6 7.5 GB current
32k 209.1 18.4 8.5 GB view
64k 184.8 13.7 10.6 GB view
Batching Results
Batch Size TG tok/s Speedup
24.2 1.00×
52.3 2.16×
72.2 2.98×
70.2 2.90×