← Back to community benchmarks

Qwen3.6-27B-oQ8-mtp

M3 Ultra (80c) · 512 GB · 8bit · 2026-05-08

Performance

16k

tokens

409.1

PP tok/s

29.1

TG tok/s

40047

TTFT (ms)

32.8

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 512 GB

GPU Cores 80

Software

oMLX v0.3.9.dev1

macOS macOS 26.4.1

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	385.9	29.2	28.8 GB	view
4k	420.0	31.9	30.3 GB	view
8k	419.8	30.9	31.3 GB	view
16k	409.1	29.1	32.8 GB	current
32k	384.2	27.5	35.8 GB	view
64k	334.0	23.6	41.8 GB	view
128k	254.1	19.6	54.2 GB	view
195k	201.4	16.1	68.9 GB	view