TARS Embodied Foundation Model Wins SAIL Award at WAIC 2026
TARS received the prestigious SAIL Award for its AWE embodied foundation model at the World Artificial Intelligence Conference in Shanghai. The model enhances robot efficiency and task execution.
TARS showcased its "Trustworthy Physical AI" vision at the World Artificial Intelligence Conference (WAIC 2026) in Shanghai. The company's AWE (AI World Engine) embodied foundation model was honored with the SAIL (Superior Al Leader) Award for technical innovation and industrial potential.
During the conference, TARS Founder and CEO Dr. Chen Yilun introduced AWE 3.5, the latest embodied-native foundation model. Trained on over one million hours of real-world industrial data, AWE 3.5 integrates action, perception, and sensing, reportedly doubling task-execution efficiency and improving complex task performance. It also utilizes a novel "pre-training + post-training" paradigm for reproducibility and scalability.
The company demonstrated AWE-powered robots performing tasks such as phone packing, backpack organization, and precision screw sorting. TARS plans to expand its training dataset to 10 million hours by the end of 2026 to further enhance generalization capabilities.
Further exhibits included a simulated automotive wiring-harness production line and the DexHand robotic hand, which performed intricate manipulations. These demonstrations highlighted TARS' strategy for physical AI, focusing on environmental understanding, production line value delivery, and high-precision task execution.