← Back to community benchmarks

Qwen3.6-27B-oQ4-mtp

M3 Ultra (80c) · 512 GB · 4bit · 2026-05-09
Performance
8k
tokens
420.7
PP tok/s
22.0
TG tok/s
19472
TTFT (ms)
23.1
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.3.8
macOS macOS 26.4.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 389.9 34.4 20.7 GB view
4k 419.8 25.9 22.1 GB view
8k 420.7 22.0 23.1 GB current
Batching Results
Batch Size TG tok/s Speedup
34.4 1.00×
40.3 1.17×
60.1 1.75×