← Back to community benchmarks

Jundot-Qwen3.6-27B-oQ8-mtp

M3 Ultra (80c) · 512 GB · 8bit · 2026-05-15

Performance

4k

tokens

415.6

PP tok/s

30.5

TG tok/s

9857

TTFT (ms)

30.3

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 512 GB

GPU Cores 80

Software

oMLX v0.3.9.dev2

macOS macOS 26.3.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	378.6	31.7	28.8 GB	view
4k	415.6	30.5	30.3 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	31.7	1.00×
2×	38.6	1.22×
4×	57.9	1.83×