← Back to community benchmarks

Qwen3.6-27B-oQ8-mtp

M1 Max (32c) · 64 GB · 8bit · 2026-05-16

Performance

8k

tokens

90.1

PP tok/s

15.9

TG tok/s

90874

TTFT (ms)

30.9

Peak mem (GB)

Hardware

Chip M1 Max (32c)

Memory 64 GB

GPU Cores 32

Software

oMLX v0.3.9.dev2

macOS macOS 26.4.1

Context 8,192