π0.5 ← Back to Talks 2026.10.20 A Vision-Language-Action Model with Open-World Generalization Paper Video Previous Next