Awesome World Models is a curated list that serves as a comprehensive resource for research and applications in world modeling.
The project addresses the challenge of navigating world modeling research across multiple disciplines and application domains. World models—systems that learn to predict and simulate future states of environments—have become increasingly important in AI, but the term carries varying definitions depending on whether researchers are working in embodied AI, autonomous driving, natural language processing, or game simulation. This repository consolidates works across these domains into a single organized collection, making it easier for researchers and practitioners to discover relevant papers, tutorials, and implementations without having to search across fragmented communities.
The resource suits anyone entering world modeling research or seeking to understand how the field applies to specific problems. It organizes content by application area—including embodied AI, autonomous driving, game simulation, and science—as well as by approach type, such as general methodologies, theory, and evaluation techniques. The collection includes links to papers, websites, and code repositories, with a badge system for quick navigation to these resources. This structure helps both newcomers find starter materials and established researchers locate domain-specific work.
The project has attracted rapid community engagement, with contributions arriving through pull requests and direct contact. The maintainers actively solicit submissions to keep the collection current and comprehensive. A contributing guide is available for those wishing to add entries. The project maintains an ongoing call for community participation to ensure the resource remains up-to-date as the field evolves.
