← Back to community benchmarks

Qwen3.6-27B-oQ4-mtp

M5 (10c) · 32 GB · 4bit · 2026-05-20
Performance
16k
tokens
143.3
PP tok/s
9.0
TG tok/s
114302
TTFT (ms)
20.6
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.9rc1
macOS macOS 26.4
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 190.3 13.4 16.6 GB view
4k 194.6 13.2 18.1 GB view
8k 178.5 11.9 19.1 GB view
16k 143.3 9.0 20.6 GB current
Batching Results
Batch Size TG tok/s Speedup
13.4 1.00×
18.3 1.37×
33.6 2.51×