iWorld-Bench: A Benchmark for Interactive World Models with a Unified Action Generation Framework

Jianjie Fang, Yingshan Lei, Qin Wan, Ziyou Wang, Yuchao Huang, Yongyan Xu, Baining Zhao, Weichen Zhang, Chen Gao, Xinlei Chen, Yong Li
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:29222-29255, 2026.

Abstract

Achieving Artificial General Intelligence (AGI) requires agents that learn and interact adaptively, with interactive world models providing scalable environments for perception, reasoning, and action. Yet current research still lacks large-scale datasets and unified benchmarks to evaluate their physical interaction capabilities. To address this, we propose iWorld-Bench, a comprehensive benchmark for training and testing world models on interaction-related abilities such as distance perception and memory. We construct a diverse dataset with 330k video clips and select 2.1k high-quality samples covering varied perspectives, weather, and scenes. As existing world models differ in interaction modalities, we introduce an Action Generation Framework to unify evaluation and design six task types, generating 4.9k test samples. These tasks jointly assess model performance across visual generation, trajectory following, and memory. Evaluating 14 representative world models, we identify key limitations and provide insights for future research. The iWorld-Bench model leaderboard is publicly available at iWorld-Bench.com.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-fang26l, title = {i{W}orld-Bench: A Benchmark for Interactive World Models with a Unified Action Generation Framework}, author = {Fang, Jianjie and Lei, Yingshan and Wan, Qin and Wang, Ziyou and Huang, Yuchao and Xu, Yongyan and Zhao, Baining and Zhang, Weichen and Gao, Chen and Chen, Xinlei and Li, Yong}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {29222--29255}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/fang26l/fang26l.pdf}, url = {https://proceedings.mlr.press/v306/fang26l.html}, abstract = {Achieving Artificial General Intelligence (AGI) requires agents that learn and interact adaptively, with interactive world models providing scalable environments for perception, reasoning, and action. Yet current research still lacks large-scale datasets and unified benchmarks to evaluate their physical interaction capabilities. To address this, we propose iWorld-Bench, a comprehensive benchmark for training and testing world models on interaction-related abilities such as distance perception and memory. We construct a diverse dataset with 330k video clips and select 2.1k high-quality samples covering varied perspectives, weather, and scenes. As existing world models differ in interaction modalities, we introduce an Action Generation Framework to unify evaluation and design six task types, generating 4.9k test samples. These tasks jointly assess model performance across visual generation, trajectory following, and memory. Evaluating 14 representative world models, we identify key limitations and provide insights for future research. The iWorld-Bench model leaderboard is publicly available at iWorld-Bench.com.} }
Endnote
%0 Conference Paper %T iWorld-Bench: A Benchmark for Interactive World Models with a Unified Action Generation Framework %A Jianjie Fang %A Yingshan Lei %A Qin Wan %A Ziyou Wang %A Yuchao Huang %A Yongyan Xu %A Baining Zhao %A Weichen Zhang %A Chen Gao %A Xinlei Chen %A Yong Li %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-fang26l %I PMLR %P 29222--29255 %U https://proceedings.mlr.press/v306/fang26l.html %V 306 %X Achieving Artificial General Intelligence (AGI) requires agents that learn and interact adaptively, with interactive world models providing scalable environments for perception, reasoning, and action. Yet current research still lacks large-scale datasets and unified benchmarks to evaluate their physical interaction capabilities. To address this, we propose iWorld-Bench, a comprehensive benchmark for training and testing world models on interaction-related abilities such as distance perception and memory. We construct a diverse dataset with 330k video clips and select 2.1k high-quality samples covering varied perspectives, weather, and scenes. As existing world models differ in interaction modalities, we introduce an Action Generation Framework to unify evaluation and design six task types, generating 4.9k test samples. These tasks jointly assess model performance across visual generation, trajectory following, and memory. Evaluating 14 representative world models, we identify key limitations and provide insights for future research. The iWorld-Bench model leaderboard is publicly available at iWorld-Bench.com.
APA
Fang, J., Lei, Y., Wan, Q., Wang, Z., Huang, Y., Xu, Y., Zhao, B., Zhang, W., Gao, C., Chen, X. & Li, Y.. (2026). iWorld-Bench: A Benchmark for Interactive World Models with a Unified Action Generation Framework. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:29222-29255 Available from https://proceedings.mlr.press/v306/fang26l.html.

Related Material