# What is World models? Definition and meaning

A learned model that predicts how an environment will change, given its current state and an action, so an agent can plan or train inside it.

## Definition

**World models**: A learned model that predicts how an environment will change, given its current state and an action, so an agent can plan or train inside it.

In short: Learn the world, then imagine.

## When it is used

Model-based reinforcement learning, robot planning, game and scene generation, and driving simulation.

## Where it fits

World models belongs to **Frontier topics** in the [Deep learning](/topics/deep-learning) topic. These are the other terms in the same group:

- **[Vision-language-action models](/glossary/vision-language-action-models)**: A neural network that takes camera input and a natural-language instruction and outputs low-level actions for a robot.

## Learn more

The full guide to **[World models](/tools/world-models)** explains how it works, when to use it and how it compares with the alternatives, with examples. The [glossary](/glossary) lists every term on the site from A to Z.

Previous term: [Word embeddings](/glossary/word-embeddings) · Next term: [XGBoost](/glossary/xgboost)
