Google has launched Gemini Robotics ER 2, an embodied reasoning model that helps robots understand their surroundings, communicate with people, and complete complex physical tasks in real time. The model achieved 57.4% accuracy in progress classification tests and 91.3% accuracy in moment finding, measuring its ability to identify precise action points. It delivers four times faster execution than larger model categories with an average timing distance of 0.96 seconds. The system supports multi-robot collaboration and includes safety features that can halt robot operation when a person enters its work area. Gemini Robotics ER 2 is available through the Gemini API and Google AI Studio.
Source: Read the original article
