← Back to community benchmarks

Qwen3.5-35B-A3B

M3 Ultra (60c) · 256 GB · 4bit · 2026-03-09
Performance
1k
tokens
1,487
PP tok/s
96.0
TG tok/s
689
TTFT (ms)
20.0
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 256 GB
GPU Cores 60
Software
oMLX v0.2.6
macOS macOS 26.3
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,487 96.0 20.0 GB current
4k 2,017 92.1 20.2 GB view
8k 2,098 87.4 20.4 GB view
16k 1,940 79.4 20.8 GB view
Batching Results
Batch Size TG tok/s Speedup
96.0 1.00×
165.8 1.73×
264.3 2.75×