Figure Drops OpenAI, Unveils Helix: Embodied AI Model Sets New Milestones
Figure AI introduces Helix, a groundbreaking VLA model for humanoid robots, achieving zero-shot collaboration, full-body control, and task generalization with natural language.
"AI Disruption" publication New Year 30% discount link.
Local GPU operation, 7B system 2 models + 80M system 1 model, allowing multiple robots to share a "brain".
After abruptly announcing the end of its collaboration with OpenAI in February, the well-known robotics startup Figure AI revealed the reason behind it this Thursday night: they have developed their own general embodied intelligence model, Helix.
Helix is a general Vision-Language-Action (VLA) model that integrates perception, language understanding, and learning control to overcome multiple long-standing challenges in robotics technology.
Helix has achieved several firsts: