“Gemini Robotics 1.5 is a multi-embodiment Vision-Language-Action (VLA) model.”