🌍Text2World: Benchmarking Large Language Models for Symbolic World Model Generation

Mengkang Hu*1, Tianxing Chen*1, Yude Zou*3, Yuheng Lei1, Qiguang Chen2, Ming Li1, Hongyuan Zhang1, Wenqi Shao3, Ping Luo†1
1The University of Hong Kong, 2Harbin Institute of Technology, 3Shanghai AI Laboratory
Paper arXiv Code Data Viewer

T2D image

Abstract

Recently, there has been growing interest in leveraging large language models (LLMs) to generate symbolic world models from textual descriptions.

Although LLMs have been extensively explored in the context of world modeling, prior studies encountered several challenges, including evaluation randomness, dependence on indirect metrics, and a limited domain scope.

To address these limitations, we introduce a novel benchmark, Text2World, based on planning domain definition language (PDDL), featuring hundreds of diverse domains and employing multi-criteria, execution-based metrics for a more robust evaluation. We benchmark current LLMs using Text2World and find that reasoning models trained with large-scale reinforcement learning outperform others.

However, even the best-performing model still demonstrates limited capabilities in world modeling.

Building on these insights, we examine several promising strategies to enhance the world modeling capabilities of LLMs, including test-time scaling, agent training, and more. We hope that Text2World can serve as a crucial resource, laying the groundwork for future research in leveraging LLMs as world models.

Table 1: Performance comparison of different LLMs on Text2World. ECk denotes the setting where models are allowed k correction attempts.

BibTeX


@misc{hu2025text2worldbenchmarkinglargelanguage,
  title={Text2World: Benchmarking Large Language Models for Symbolic World Model Generation}, 
  author={Mengkang Hu and Tianxing Chen and Yude Zou and Yuheng Lei and Qiguang Chen and Ming Li and Hongyuan Zhang and Wenqi Shao and Ping Luo},
  year={2025},
  eprint={2502.13092},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2502.13092}, 
}