Simulac
Accelerate physical AI (VLA) evaluation
Start new thread
Tektonian

Simulac - Accelerate physical AI (VLA) evaluation

by•
Running VLA benchmarks requires setup and dependencies. This launch removes that overhead with a no-setup execution layer for one-line benchmark execution, enabling parallel, cross-benchmark evaluation at scale.

Add a comment

Replies

Best
Tektonian
Maker
📌
Happy to share this. In our experience, running robot benchmarks often ends up being more about setup and debugging than evaluation itself. This was built to remove that overhead. Curious how others are dealing with this today: - Are you running everything locally? - Any workflows for parallel execution or scaling experiments? Would be great to hear what’s actually working in practice.