← Back to community benchmarks

Qwen3.6-27B-UD

M5 Max (40c) · 128 GB · 4bit · 2026-04-30
Performance
8k
tokens
727.1
PP tok/s
20.2
TG tok/s
11267
TTFT (ms)
27.3
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.6
macOS macOS 26.4.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 781.0 21.3 25.2 GB view
4k 835.7 20.9 26.6 GB view
8k 727.1 20.2 27.3 GB current
16k 668.8 18.9 28.5 GB view
32k 609.9 17.5 31.0 GB view
Batching Results
Batch Size TG tok/s Speedup
21.3 1.00×
23.6 1.11×
37.8 1.77×