← Back to community benchmarks

Qwen3.6-27B-oQ4e-mtp

M5 Max (40c) · 128 GB · 4bit · 2026-07-13
User commentjundot/Qwen3.6-27B-oQ4e-mtp
Performance
16k
tokens
732.6
PP tok/s
28.6
TG tok/s
22366
TTFT (ms)
20.6
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.5.1
macOS macOS 26.5.2
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 855.1 31.0 16.4 GB view
4k 879.1 30.6 17.9 GB view
8k 869.8 30.0 18.8 GB view
16k 732.6 28.6 20.6 GB current
32k 678.0 27.3 23.9 GB view
64k 560.6 24.0 27.6 GB view
128k 372.4 18.3 37.9 GB view
Batching Results
Batch Size TG tok/s Speedup
31.0 1.00×
52.5 1.69×
77.6 2.50×