
How do we make robot policy evaluation reproducible across labs? Announcing OpenArm 02: a fully open-source dual arm platform, with a unified setup for reproducible evaluation: controlled tasks, backgrounds, lighting, and camera angles. ✅ Fully open-source, so anyone can replicate the setup anywhere ✅ AutoEval: 24/7 real-world evaluation with minimal human intervention (building on work by @zhiyuan_zhou_) ✅ Fair side-by-side policy comparison OpenArm is the open foundation to turn isolated results into shared progress. Releasing soon. Stay tuned.















