BEHAVIOR-1K

'BEHAVIOR-1K' is a benchmark that evaluates the performance of Vision-Language-Action models in embodied tasks.

Papers using BEHAVIOR-1K (6)

Make Your VLA More Robust Without More Data By Interleaving Motion Planning2026SERF: Spatiotemporal Environment and Robot Feature Map for Long-Horizon Mobile Manipulation2026StarVLA: A Lego-like Codebase for Vision-Language-Action Model Developing2026Openpi Comet: Competition Solution For 2025 BEHAVIOR Challenge2025LookPlanGraph: Embodied Instruction Following Method with VLM Graph Augmentation2025MORE: Mobile Manipulation Rearrangement Through Grounded Language Reasoning2025