← Back to community benchmarks

Qwen3.6-27B-oQ8-mtp

M3 Ultra (80c) · 512 GB · 8bit · 2026-05-24

Performance

32k

tokens

374.7

PP tok/s

25.6

TG tok/s

87442

TTFT (ms)

35.0

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 512 GB

GPU Cores 80

Software

oMLX v0.3.9

macOS macOS 26.5

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	329.3	29.8	28.8 GB	view
4k	405.4	26.0	30.3 GB	view
32k	374.7	25.6	35.0 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	29.8	1.00×
2×	36.1	1.21×
4×	56.1	1.88×