The Google DeepMind team has unveiled two cutting - edge AI models tailored for the real - world environment: Gemini Robotics 1.5 and Gemini Robotics - ER 1.5. Their overarching goal is to forge physical agents. These models are built on a 'brain - body' synergistic framework. In this setup, Gemini Robotics - ER 1.5 functions as the 'higher - order brain', taking charge of planning and decision - making processes. On the other hand, Gemini Robotics 1.5 plays the role of the 'body', converting instructions into precise robotic motion commands. Gemini Robotics 1.5 is equipped with the remarkable ability to 'think before acting'. It also showcases outstanding cross - robot morphology learning abilities, which means it can adapt and learn from different robotic forms. At present, Google is vigorously working on safety and alignment techniques. Both of these innovative models are already available to developers and a select group of partners.