Google DeepMind unveils Gemini Robotics 2 for whole-body robot control

Google DeepMind announced Gemini Robotics 2, the intelligence layer powering the next generation of adaptable robots. The update unlocks intelligent whole-body control, advanced dexterity, and multi-robot collaboration, letting humanoids walk, crouch, stretch, and manipulate objects to clean up a cluttered room. The models can even run locally on-device and adapt to entirely new robotic bodies in just a few hours.

The release consists of three models: Gemini Robotics 2, the most advanced vision-language-action model (VLA), capable of controlling full humanoids from feet to fingertips; Gemini Robotics ER 2, an embodied reasoning model that acts as the robot's high-level brain, communicating with humans and planning multi-step tasks; and Gemini Robotics On-Device 2, an efficient VLA optimized to run locally on robotic devices. ER 2 is available on Google AI Studio and in private preview on the Gemini Enterprise Agent Platform.

DeepMind also introduced multi-robot collaboration, enabling different robot types to communicate and work together on workflows a single robot cannot handle alone. On safety, the team released ASIMOV-Agentic, a new benchmark for agentic safety orchestration and uncertainty resolution, and says ER 2 is its safest robotics model to date in safety constraint following and human proximity benchmarks. Development partners include Apptronik, Boston Dynamics, and Agile Robots.