close

BEHAVIOR-1K is the first simulation benchmark grounded in real human needs. Based on extensive surveys asking "what do you want robots to do for you?", we present 1,000 household activities that people actually spend time on and want help with.

The benchmark introduces long-horizon mobile manipulation challenges in realistic environments, providing a crucial tool to help bridge the gap between current research and real-world, human-centric applications.

1,000 activities, instantiated in 50 fully interactive scenes with 10,000+ objects

OmniGibson: Realism in perception, physics, and interactions

Fluids
Deformables
Transparency
Thermal effects
Interact with fluids
Grasp cloth
Grasp rigid objects
Transitions

Core Team

Wensi Ai Wensi Ai
Stef Ren Stef Ren
Cem Gokmen Cem Gokmen
Yalcin Tur Yalcin Tur
Minyeong Kim Minyeong Kim
Brenda Chen Brenda Chen
Andi Xu Andi Xu
Chengshu Li Chengshu Li
Josiah Wong Josiah Wong
Hang Yin Hang Yin
Sanjana Srivastava Sanjana Srivastava
Raven Huang Raven Huang
Ruohan Zhang Ruohan Zhang
Jiajun Wu Jiajun Wu
Li Fei-Fei Li Fei-Fei

Supporting Team

Evans Han Evans Han
Lynn Jin Lynn Jin
Yunfan Jiang Yunfan Jiang
Wenlong Huang Wenlong Huang
Weiyu Liu Weiyu Liu
Manling Li Manling Li
Yunzhu Li Yunzhu Li
Roberto Martín-Martín Roberto Martín-Martín
Silvio Savarese Silvio Savarese
Hyowon Gweon Hyowon Gweon
Karen Liu Karen Liu

Contact

For technical support, please join our Discord

For other purposes, please email behavior-contact@googlegroups.com