Gemini Robotics is an advanced vision-language-action model developed by Google DeepMind [1] in partnership with Apptronik. [2] It is based on the Gemini 2.0 large language model. [3] It is tailored for robotics applications and can understand new situations. [4] [5] There is a related version called Gemini Robotics-ER, which stands for embodied reasoning. [3] The two models were launched on March 12, 2025. [5]
On June 24, 2025, Google DeepMind released Gemini Robotics On-Device, a variant designed and optimized to run locally on robotic devices. [6]
Access to Gemini Robotics models is currently restricted to trusted testers, including Agile Robots, Agility Robots, Boston Dynamics, and Enchanted Tools. [2]