← Back to community benchmarks

Qwen3.5-9B

M3 Max (40c) · 128 GB · 4bit · 2026-03-10
Performance
1k
tokens
758.9
PP tok/s
69.3
TG tok/s
1349
TTFT (ms)
6.7
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.2.6
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 758.9 69.3 6.7 GB current
4k 763.4 66.3 7.0 GB view
8k 749.8 63.8 7.1 GB view
16k 675.1 58.7 7.6 GB view
32k 628.8 47.8 8.7 GB view
Batching Results
Batch Size TG tok/s Speedup
69.3 1.00×
123.9 1.79×
132.2 1.91×
132.8 1.92×