← Back to community benchmarks

Qwen3.6-27B-UD

M3 Max (40c) · 48 GB · 4bit · 2026-04-25
Performance
32k
tokens
168.8
PP tok/s
10.1
TG tok/s
194112
TTFT (ms)
31.0
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 48 GB
GPU Cores 40
Software
oMLX v0.3.6
macOS macOS 26.4.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 206.9 14.4 25.2 GB view
4k 205.5 14.2 26.6 GB view
8k 196.8 13.8 27.3 GB view
16k 202.1 11.7 28.5 GB view
32k 168.8 10.1 31.0 GB current
Batching Results
Batch Size TG tok/s Speedup
14.4 1.00×
15.6 1.08×
25.1 1.74×