Google has introduced Gemini Robotics and Gemini Robotics-ER, AI models designed to enhance robots' interaction with the physical world.
Based on Gemini 2.0, these models enable robots to understand their surroundings and adapt in real-time.
Gemini Robotics combines AI with physical actions, while Gemini Robotics-ER offers advanced spatial understanding, allowing developers to integrate their own programs.
Google's CEO, Sundar Pichai, highlighted the potential of these models for robots to respond and adapt to their environments.
Gemini Robotics was tested on data from Google's bi-arm robotics platform, ALOHA 2, but can be specialized for complex use cases such as Apptronik's Apollo robot. These models are designed for robots of all form factors including humanoids and other types used in factories and warehouses.
This launch follows robotics startup Figure AI's exit from its collaboration agreement with OpenAI after it made an internal breakthrough in AI for robots.
Google's Project Astra, an experimental AI assistant, complements Gemini Robotics by enabling real-time recognition and response to physical objects.