← Back to community benchmarks

Qwen3.6-27B-UD

M3 Ultra (80c) · 512 GB · 4bit · 2026-04-29

Performance

16k

tokens

301.3

PP tok/s

68.2

TG tok/s

54377

TTFT (ms)

33.2

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 512 GB

GPU Cores 80

Software

oMLX v0.3.8rc1

macOS macOS 26.4.1

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	378.9	122.3	28.9 GB	view
4k	371.2	90.0	30.7 GB	view
8k	344.1	79.8	31.5 GB	view
16k	301.3	68.2	33.2 GB	current
32k	224.3	47.1	36.1 GB	view