← Back to community benchmarks

Qwen3.6-27B-MTPLX-Optimized-Speed

M2 Max (38c) · 96 GB · 4bit · 2026-05-23
Performance
4k
tokens
96.3
PP tok/s
14.2
TG tok/s
42535
TTFT (ms)
17.3
Peak mem (GB)
Hardware
Chip M2 Max (38c)
Memory 96 GB
GPU Cores 38
Software
oMLX v0.3.9
macOS macOS 26.5
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 105.1 15.3 15.8 GB view
4k 96.3 14.2 17.3 GB current
8k 93.9 13.2 18.3 GB view
16k 91.8 12.8 19.8 GB view
Batching Results
Batch Size TG tok/s Speedup
15.3 1.00×
18.4 1.20×
22.2 1.45×
20.6 1.35×