← Back to community benchmarks

Qwen3.6-27B-oQ8-mtp

M3 Ultra (80c) · 512 GB · 8bit · 2026-05-20
Performance
4k
tokens
421.2
PP tok/s
31.2
TG tok/s
9724
TTFT (ms)
30.3
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.3.9rc1
macOS macOS 26.5
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 387.3 32.6 28.8 GB view
4k 421.2 31.2 30.3 GB current
8k 420.6 30.7 31.3 GB view
16k 409.9 29.5 32.8 GB view
32k 384.5 28.0 35.8 GB view
64k 335.1 24.2 41.8 GB view