OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments
Tianbao Xie, Danyang Zhang, Jixuan Chen, Xiaochuan Li, Siheng Zhao, Ruisheng Cao, Toh Jing Hua, Zhoujun Cheng, Dongchan Shin, Fangzhou Lei, Yitao Liu, Yiheng Xu, Ting-Wei Kuo, Tao Yu, Pan Lu, Lei Li, Ying Sheng, Bo Li, Yuke Zhu, Percy Liang, Didi Zhu, Zhiyong Wu, Zhilin Yang
Abstract
A benchmark of 369 real tasks across real operating systems (Ubuntu/Windows/macOS), requiring agents to control GUIs like a human to complete open-ended computer tasks.
Cite this entry
GB/T 7714-2015
Tianbao Xie, Danyang Zhang, Jixuan Chen, et al. OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments[EB/OL]. NeurIPS 2024, 2024(2024-04-11)[2026-09-16]. https://arxiv.org/abs/2404.07972.
BibTeX
@misc{xie2024,
author = {Tianbao Xie and Danyang Zhang and Jixuan Chen and Xiaochuan Li and Siheng Zhao and Ruisheng Cao and Toh Jing Hua and Zhoujun Cheng and Dongchan Shin and Fangzhou Lei and Yitao Liu and Yiheng Xu and Ting-Wei Kuo and Tao Yu and Pan Lu and Lei Li and Ying Sheng and Bo Li and Yuke Zhu and Percy Liang and Didi Zhu and Zhiyong Wu and Zhilin Yang},
title = {OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments},
year = {2024},
organization = {NeurIPS 2024},
howpublished = {\url{https://arxiv.org/abs/2404.07972}},
} This entry is part of the AgentNet Observer library. Attribute with a link to this page when quoting.