High-Level Robot Brains

Gemini Robotics 2 Expands Physical AI into Whole-Body Motions

Gemini Robotics proved Gemini's understanding of multiple types of input and how to use them to take real-world action, and with Gemini Robotics 2, DeepMind (the artificial intelligence research lab owned by Google's parent company) is bringing whole-body intelligence to robots. As it stands, robots are mostly pre-programmed or teleoperated, and can only carry out short, repetitive tasks that they've been taught. Gemini Robotics 2, the intelligence layer powering the next generation of adaptable robots, helps robots reason, take on a broad range of tasks, walk, crouch, stretch, and manipulate objects, and even work collaboratively with other robots to get a job done in less time.

Alongside Gemini Robotics 2, DeepMind is introducing Gemini Robotics ER 2, an advanced embodied reasoning model, and Gemini Robotics On-Device 2, a vision-language-action model optimized to run locally on robotic devices.

Open at source ↗