← Back to community benchmarks

Qwen3.5-27B

M4 Pro (20c) · 64 GB · 4bit · 2026-03-10
Performance
32k
tokens
111.9
PP tok/s
11.4
TG tok/s
292862
TTFT (ms)
20.1
Peak mem (GB)
Hardware
Chip M4 Pro (20c)
Memory 64 GB
GPU Cores 20
Software
oMLX v0.2.6
macOS macOS 26.2
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 126.5 15.5 16.6 GB view
4k 124.8 15.1 17.0 GB view
8k 123.9 14.7 17.4 GB view
16k 119.0 13.7 18.3 GB view
32k 111.9 11.4 20.1 GB current
Batching Results
Batch Size TG tok/s Speedup
15.5 1.00×
26.9 1.74×