Skip to main content
The Keyword

How we built the new family of Gemini Robotics models

A still of a white and black robot in a kitchen packing a blue lunchbox

Carolina says witnessing the slam dunk was a “wow” moment.

Gemini Robotics-ER excels at embodied reasoning capabilities, including detecting objects and pointing at object parts, finding corresponding points and detecting objects in 3D.

This is a collage of visualizations showcasing these capabilities. Top left: 2D object detection, top right: pointing, bottom left: multi-view correspondence, bottom right: 3d object detection.

The models adapt to different embodiments, able to perform tasks like packing a lunchbox or wiping a whiteboard in different forms.

Four images of robots performing actions. In the top left, a humanoid robot is packing a lunch, in the top right a small arm can be seen picking up a snap pea from a tupperware container, in the bottom left two large white arms ready for a task on a bench, and in the bottom right a black pincer hand holds a whiteboard eraser atop a whiteboard.

Let’s stay in touch. Get the latest news from Google in your inbox.

Subscribe