← Back to community benchmarks

Qwen3.5-35B-A3B

M3 Ultra (60c) · 96 GB · 4bit · 2026-04-15
User commentDflash + all parameters enabled
Performance
16k
tokens
2,011
PP tok/s
83.4
TG tok/s
8148
TTFT (ms)
40.0
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 96 GB
GPU Cores 60
Software
oMLX v0.3.5
macOS macOS 26.3
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,809 398.3 38.7 GB view
4k 1,966 100.2 39.1 GB view
8k 2,132 92.5 39.4 GB view
16k 2,011 83.4 40.0 GB current
32k 1,745 72.2 41.4 GB view
64k 1,327 55.7 44.1 GB view
128k 841.4 25.7 49.4 GB view
195k 603.4 5.6 41.5 GB view