Google DeepMind Unveils Gemini Robotics 2, Giving Humanoid Robots Whole-Body Intelligence and Multi-Robot Collaboration
Summary
Google DeepMind unveils Gemini Robotics 2, a groundbreaking AI system giving humanoid robots whole-body intelligence, multi-robot collaboration, and built-in safety features across three specialized models.
Key Points
- Google DeepMind is introducing Gemini Robotics 2, a next-generation AI system that brings whole-body intelligence to robots, enabling humanoids to walk, crouch, manipulate objects, and collaborate with other robots to complete complex real-world tasks.
- The release includes three specialized models: Gemini Robotics 2 for full-body motor control and dexterous manipulation, Gemini Robotics ER 2 for high-level agentic reasoning and multi-robot collaboration, and Gemini Robotics On-Device 2 for efficient local operation that adapts to new robot bodies in just a few hours with minimal data.
- Safety remains a core priority, with the introduction of ASIMOV-Agentic, a new benchmark for agentic safety that measures a robot's ability to refuse unsafe commands, predict task feasibility, and detect human proximity to trigger protective stops.