Gemini Robotics 1.5: Thinking while acting

Channel:

Google DeepMind

Subscribers:

663,000

Published on September 25, 2025 3:53:55 PM ● Video Link: https://www.youtube.com/watch?v=eDyXEh8XqjM

Duration: 0:00

30,732 views

1,299

Previously, robots could complete one tasks per instruction. Now, they can solve longer multi-step challenges. This means they can generate an internal sequence of reasoning and analysis in natural language to perform tasks that require multiple steps or require a deeper semantic understanding.

Learn more at https://deepmind.google/models/gemini-robotics/
___

Subscribe to our channel    / @googledeepmind  
Find us on X https://twitter.com/GoogleDeepMind
Follow us on Instagram https://instagram.com/googledeepmind
Add us on Linkedin https://www.linkedin.com/company/deepmind/