# World Models

A **World Model** is a generative AI that learns an internal, predictive model of its environment. Think of it as a robot's "imagination." It allows the robot to simulate "what would happen if...?" without having to perform the action in the real world. By understanding the rules, physics, and dynamics of its surroundings, a world model can predict future states from current observations and potential actions.

Prominent examples like **DeepMind's Genie 2** and models developed by **Worldlabs** can generate high-fidelity, interactive simulation environments from multi-view imagery and video data. A robot can use these internal simulations to:

* Train and learn new skills in a safe, virtual space.
* Plan complex, multi-step tasks by exploring the outcomes of different action sequences.
* Anticipate the actions of other agents (e.g., humans or other robots).

Essentially, world models give robots a form of common-sense physical reasoning inside realistic environments, drastically improving their planning and decision-making capabilities.

<br>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.overthereality.ai/over-wiki/physical-ai/physical-ai-foundation-models-for-robotics/world-models.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.