Google DeepMind launches Gemini Robotics 2 for whole-body robot control

Google DeepMind introduced Gemini Robotics 2, a suite of three models enabling intelligent whole-body control of humanoid robots from feet to fingertips, advanced dexterity with multi-finger hands and grippers, multi-robot collaboration, and rapid on-device adaptation to new robot embodiments in a few hours. The vision-language-action model controls full humanoids such as Apptronik Apollo 2, while the embodied reasoning model handles multi-step planning lasting several minutes and the on-device variant runs locally. Gemini Robotics ER 2 is available in Google AI Studio and private preview, with the other models offered to early-access partners, advancing general-purpose physical AI capabilities.