← Back to community benchmarks

Qwen3.5-9B

M4 (10c) · 16 GB · 4bit · 2026-03-10
Performance
16k
tokens
201.5
PP tok/s
18.5
TG tok/s
81299
TTFT (ms)
6.7
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 213.6 20.9 5.7 GB view
4k 223.4 21.3 5.9 GB view
8k 219.0 20.3 6.2 GB view
16k 201.5 18.5 6.7 GB current
Batching Results
Batch Size TG tok/s Speedup
20.9 1.00×
42.9 2.05×
50.3 2.41×
51.0 2.44×