← Back to community benchmarks

Qwen3.6-27B-MTPLX-Optimized-Speed

M2 Ultra (76c) · 192 GB · 4bit · 2026-05-16

Performance

4k

tokens

245.2

PP tok/s

8.9

TG tok/s

16706

TTFT (ms)

17.3

Peak mem (GB)

Hardware

Chip M2 Ultra (76c)

Memory 192 GB

GPU Cores 76

Software

oMLX v0.3.9.dev2

macOS macOS 14.8.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	119.7	6.2	15.8 GB	view
4k	245.2	8.9	17.3 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	6.2	1.00×
2×	30.8	4.97×
4×	25.9	4.18×