September,2026,The WALL-B-powered TwinDEX system enables long-horizon dexterous manipulation.

September 2026,The WALL- B powered high-DoF dexterous hand ,enable to learn a wide range of complex, precise manipulation tasks quickly.

June 2026 | WALL-B-powered, stable and lightweight hardware: Sorting irregular parcels on a live logistics line and separating small NC items with a consistent orientation.

Further unlocked the potential of high-DoF dexterous hands in December 2025.

Completed the world’s first end-to-end VLA-based mobile operation spanning outdoor and indoor real open scenarios in October 2025.

Pioneered truly autonomous operation of high-degree-of-freedom (DoF) dexterous hands in August 2025.

Made further breakthroughs in reinforcement learning, few-shot learning, and cross-task transfer capabilities in August 2025.

Completed autonomous planning and execution of a full set of long-sequence complex tasks with the QUANTA X1 Multi-Robot Intelligent Swarm in July 2025.

Executed real-world long-horizon tasks in open scenarios in May 2025

Achieved industry-leading breakthroughs in Embodied Chain of Thought (CoT) and related domains in March 2025

Achieved breakthroughs in semantic navigation and semantic understanding-based manipulation. WALL-A exhibited zero-shot generalization in unseen task scenarios in January 2025

Developed WALL-A, the world’s leading end-to-end unified embodied foundation model with the largest parameter scale in October 2024

Demonstrated few-shot learning and cross-task transfer in June 2024.

Released China's earliest Embodied Intelligence Foundation Model in April 2024.