← Back to community benchmarks

Qwen3___5-35B-A3B

M3 Ultra (80c) · 512 GB · unknown · 2026-03-10
Performance
1k
tokens
517.8
PP tok/s
2.9
TG tok/s
1978
TTFT (ms)
72.6
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.2.6
macOS macOS 15.3
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 517.8 2.9 72.6 GB current
4k 777.6 2.9 72.7 GB view
Batching Results
Batch Size TG tok/s Speedup
2.9 1.00×
5.6 1.93×
10.9 3.76×