Helix by Figure AI — Vision-language-action model

Figure's AI model for humanoid perception and whole-body manipulation, with the Helix 2.5 update demonstrating zero-shot generalization across unseen homes in 2