← Back to community benchmarks

Qwen3.6-27B-UD

M5 Max (40c) · 128 GB · 4bit · 2026-07-01
Performance
195k
tokens
137.6
PP tok/s
9.8
TG tok/s
1453698
TTFT (ms)
59.2
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.4.4
macOS macOS 26.5.1
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 575.9 22.1 25.2 GB view
4k 827.1 21.7 26.6 GB view
8k 793.0 21.6 27.6 GB view
16k 709.8 20.5 29.1 GB view
32k 631.0 19.4 32.1 GB view
64k 507.8 15.3 38.2 GB view
128k 309.8 12.0 50.6 GB view
195k 137.6 9.8 59.2 GB current
Batching Results
Batch Size TG tok/s Speedup
22.1 1.00×
18.4 0.83×
29.7 1.34×