Gemini Robotics 2 from Google DeepMind takes a leap forward with the ability to perform complex physical tasks thanks to its enhanced whole-body intelligence. The new platform, released just over a year after the initial Gemini Robotics debut, enables robots to smoothly interpret their surroundings, understand language commands, and execute coordinated actions over longer timeframes without human intervention.
One standout feature is improved dexterity robots can now manipulate objects and interact with their environment in ways that previous models could not handle. The system also supports multi-robot collaboration, allowing different robot types such as bipeds and humanoids to work together smoothly. Google’s partnerships with companies like Apptronik, Boston Dynamics, and Agile Robots demonstrate this capability in action.
Developers gain access to Gemini Robotics ER 2 through Google AI Studio, inviting third-party innovation in embodied robotic intelligence. This aligns with Google's vision of 'physical AGI,' where robots achieve general intelligence in the physical world.
