← Back to community benchmarks

Qwen3.5-35B-A3B

M3 Ultra (60c) · 96 GB · 8bit · 2026-04-15

Performance

16k

tokens

1,992

PP tok/s

69.3

TG tok/s

8224

TTFT (ms)

37.1

Peak mem (GB)

Hardware

Chip M3 Ultra (60c)

Memory 96 GB

GPU Cores 60

Software

oMLX v0.3.4

macOS macOS 26.4

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,468	81.6	35.4 GB	view
4k	2,057	80.7	36.2 GB	view
8k	2,100	75.9	36.5 GB	view
16k	1,992	69.3	37.1 GB	current
32k	1,723	58.7	38.5 GB	view
64k	1,310	43.7	41.2 GB	view
128k	834.9	29.9	46.5 GB	view