TypeSafe AI
Typed AI decisions for classification, scoring, and automated workflows
Gemini Robotics
AI models for robotic perception, planning, and physical action
Meet Gemini Robotics
Gemini Robotics is a family of AI models from Google DeepMind for helping robots understand and act in the physical world. Its vision-language-action models connect camera input and spoken or written instructions to robot movements, while embodied reasoning models support spatial understanding and planning across multiple steps. An on-device model is designed to run directly on robot hardware.
The models are intended for robotics developers working on manipulation, navigation, and tasks that require adapting to unfamiliar objects or environments. The website presents demonstrations and technical information, with access routes that vary by model: developers can explore embodied reasoning through Google AI Studio, while some robot-control models use a trusted tester program.
Discussion
Tried it? Share a useful thought or ask a question.
Good conversations start with a hello.
Sign in to commentProduct details
- Website
- deepmind.google
- Status
- In beta
- Category
- Artificial intelligence
- Listed
Explore related topics
Similar products
PrismML
Compact AI models for local assistants and applications
Telegraph Protocol
Find and route requests to AI models, APIs, and data ranked by task
Ai2
Open AI models, datasets, and tools for research and development
NeuroBlock
Create custom AI models and connect document knowledge to applications and internal workflows.
Morph
Add inference, code search, and edit application to AI agent workflows.