← Back to community benchmarks

Qwen3.6-27B-MXFP4-MTP

M2 Max (38c) · 64 GB · mxfp4 · 2026-06-04
Performance
8k
tokens
86.5
PP tok/s
12.1
TG tok/s
94656
TTFT (ms)
17.1
Peak mem (GB)
Hardware
Chip M2 Max (38c)
Memory 64 GB
GPU Cores 38
Software
oMLX v0.4.0
macOS macOS 26.5
Context 8,192