Google DeepMind presents Gemini Robotics 1.5, powering an era of physical agents. The vision-language-action model lets robots perceive, plan, think, use tools, and act to solve complex multi-step real-world tasks.