← Back to community benchmarks

Qwen3.6-27B-oQ6-mtp

M5 Pro (20c) · 48 GB · 6bit · 2026-05-20
Performance
4k
tokens
441.5
PP tok/s
19.2
TG tok/s
9277
TTFT (ms)
24.4
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 48 GB
GPU Cores 20
Software
oMLX v0.3.9rc1
macOS macOS 26.4.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 394.8 19.9 23.0 GB view
4k 441.5 19.2 24.4 GB current
8k 437.6 18.7 25.4 GB view
16k 423.8 18.1 26.9 GB view
32k 390.9 16.6 29.9 GB view
64k 324.2 14.1 36.0 GB view
Batching Results
Batch Size TG tok/s Speedup
19.9 1.00×
39.5 1.98×
76.7 3.85×