Gemini Robotics 2.5 accelerates real-world general dexterity

Google DeepMind has introduced Gemini Robotics 2.5, its most capable vision-language-action model to date, designed to improve robot dexterity, reliability, and safety. In a real-world proof-of-concept on Mercedes-Benz production lines, the model helped Apptronik’s Apollo 2 humanoid pick and pack automotive parts with a 96% success rate. The release also includes advanced inference-time reasoning guidance for hazard avoidance and features accelerated development powered by Gemini 4 Argon. Open article →