← Back to community benchmarks

Qwen3.6-35B-A3B-Abliterated-Heretic

M3 Ultra (60c) · 256 GB · 4bit · 2026-04-26

Performance

8k

tokens

2,114

PP tok/s

85.2

TG tok/s

3876

TTFT (ms)

55.4

Peak mem (GB)

Hardware

Chip M3 Ultra (60c)

Memory 256 GB

GPU Cores 60

Software

oMLX v0.3.7

macOS macOS 26.4.1

Context 8,192

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,510	93.6	54.3 GB	view
4k	2,072	90.0	55.0 GB	view
8k	2,114	85.2	55.4 GB	current
16k	1,996	78.1	56.0 GB	view
32k	1,728	67.3	57.4 GB	view
64k	1,315	52.2	60.0 GB	view
128k	842.1	35.8	65.4 GB	view
195k	608.2	22.8	71.1 GB	view