← Back to community benchmarks

Qwen3.5-27B-heretic

M3 Ultra (60c) · 96 GB · 8bit · 2026-03-11

Performance

4k

tokens

316.9

PP tok/s

21.6

TG tok/s

12925

TTFT (ms)

29.5

Peak mem (GB)

Hardware

Chip M3 Ultra (60c)

Memory 96 GB

GPU Cores 60

Software

oMLX v0.2.7

macOS macOS 15.7.4

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	313.0	22.7	29.1 GB	view
4k	316.9	21.6	29.5 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	22.7	1.00×
2×	38.7	1.70×
4×	49.3	2.17×