← Back to community benchmarks

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

M2 Ultra (60c) · 64 GB · 4bit · 2026-05-11

Performance

4k

tokens

182.2

PP tok/s

31.3

TG tok/s

22479

TTFT (ms)

17.8

Peak mem (GB)

Hardware

Chip M2 Ultra (60c)

Memory 64 GB

GPU Cores 60

Software

oMLX v0.3.9.dev1

macOS macOS 26.3.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	175.5	32.9	16.4 GB	view
4k	182.2	31.3	17.8 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	32.9	1.00×
2×	40.9	1.24×
4×	81.7	2.48×