Figure AI has announced a breakthrough for its humanoid robot, Figure 02, which can now load a dishwasher with the dexterity of a human. This milestone was achieved through the company’s new Vision Language Action (VLA) model, a system designed to help robots learn skills by observing people.
The robot demonstrates finger-level precision, carefully stacking plates, glasses, and utensils within the tight clearances of a dishwasher. Unlike traditional robots that rely on pre-programmed sequences, Figure 02 adapts dynamically to different dish sizes and placements, showing flexibility similar to a human operator.
The advancement is powered by Helix, Figure’s large AI model introduced earlier this year. Helix enables Figure 02 to acquire new abilities by simply watching demonstrations instead of requiring extensive manual coding. The system also allows the robot to recover gracefully from mistakes, such as minor bumps or misplaced dishes, ensuring smoother operation in real-world environments.
Since Helix launched in February, Figure 02 has been steadily expanding its skill set, taking on household tasks that go beyond simple motion execution. By combining vision, language understanding, and action control, the robot can interpret instructions, learn from context, and perform multi-step activities with growing reliability.
Analysts view this development as a significant step toward practical humanoid assistance in everyday life. Tasks like dishwashing require delicate handling and precise coordination, challenges that robots have historically struggled with. Figure’s achievement suggests humanoid robots may soon become capable helpers in homes, workplaces, and care environments.
As AI-powered robotics continues to evolve, Figure’s progress with Figure 02 highlights how learning from human demonstration may accelerate the path toward general-purpose household robots. The company’s next challenge will be scaling these abilities for broader adoption and proving the technology’s value beyond the lab.
