← Back to community benchmarks

Qwen3.6-27B-oQ4-mtp

M5 Max (40c) · 128 GB · 4bit · 2026-05-10
Performance
195k
tokens
308.5
PP tok/s
15.6
TG tok/s
648238
TTFT (ms)
56.8
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.9.dev1
macOS macOS 26.4.1
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 796.1 46.0 16.8 GB view
4k 912.0 46.2 18.2 GB view
8k 858.2 43.9 19.2 GB view
16k 791.3 42.3 20.7 GB view
32k 710.9 37.9 23.7 GB view
64k 594.2 29.6 29.8 GB view
128k 412.4 21.7 42.1 GB view
195k 308.5 15.6 56.8 GB current
Batching Results
Batch Size TG tok/s Speedup
46.0 1.00×
45.4 0.99×