titleWorkflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields
authorsLiya Zhu, Jingzhe Ding, Jian Zhang, Jianbo Xue, Shihao Liang, Ge Zhang, Xiang Gao, Qingshui Gu, Mailun Gao, Huimin Che, Yan Zhao, Peiheng Zhou, Haojun Wang, Chaobo Xian, Lili Le, Chi Wu, Yiwei Liu, Shengda Long, Jiale Yang, Fangzhi Xu, Sijin Wu, Haodong Duan, Yi Zhu, Chao He, Zhaojian Li, Minchao Wang, Huan Zhou, Jiani Hou, Chuqian Yu, Weiran Shi, Hongwan Gao, Jiamin Chen, Guanhong Chen, Tingqin Luo, Kaiyuan Zhang, Zhixin Yao, Qing Hua, Yuhao Jiang, Jin Chen, Pu Chen, Zhenyu Hu, Xingyu Li, Zhengxuan Jiang, Meng Cao, Tianfeng Long, Haozhe Wang, Mingzhang Wang, Yichen Zhang, Yiming Dai, Chenchen Zhang, Jiaying Wang, Zhiyong Wu, Shen Yan, Yujia Qin, Wenhao Huang, Zaiyuan Wang, Xiaolong Chang
date2026-06-09
venuearXiv (preprint, not peer reviewed)
idarXiv:2606.11042
affiliationByteDance Seed
upvotes24 (Hugging Face Daily Papers, 2026-11-01)
linksabs pdf hf
abstractRecent years have witnessed the rapid evolution of AI agents toward handling increasingly complex, real-world tasks. However, existing benchmarks rarely evaluate whether agents can operate graphical user interfaces to complete long-horizon, high-value professional workflows across diverse domains. Current GUI benchmarks still predominantly focus on general-purpose software, relatively simple applications, and short-horizon tasks, leaving it largely unknown whether modern agents can follow user instructions to autonomously operate domain-specific professional software and accomplish economically valuable work in an end-to-end manner. To bridge this gap, we introduce Workflow-GYM, a benchmark for long-horizon GUI tasks centered on professional domains and specialized software environments. Through extensive experiments on state-of-the-art models, we find that even the strongest models achieve only slightly above 30% success rates, highlighting that professional long-horizon GUI workflows remain highly challenging for current GUI agents. Further analysis reveals that current agents struggle to maintain long-horizon workflow consistency, frequently exhibiting workflow stage omission, error propagation, objective drift, and insufficient understanding of professional software environments. Our findings provide important insights into the limitations of current agent systems and suggest key directions for the next generation of GUI-agent research.

cite

@misc{zhu2026workflow,
  title={Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields},
  author={Liya Zhu and Jingzhe Ding and Jian Zhang and Jianbo Xue and Shihao Liang and Ge Zhang and Xiang Gao and Qingshui Gu and Mailun Gao and Huimin Che and Yan Zhao and Peiheng Zhou and Haojun Wang and Chaobo Xian and Lili Le and Chi Wu and Yiwei Liu and Shengda Long and Jiale Yang and Fangzhi Xu and Sijin Wu and Haodong Duan and Yi Zhu and Chao He and Zhaojian Li and Minchao Wang and Huan Zhou and Jiani Hou and Chuqian Yu and Weiran Shi and Hongwan Gao and Jiamin Chen and Guanhong Chen and Tingqin Luo and Kaiyuan Zhang and Zhixin Yao and Qing Hua and Yuhao Jiang and Jin Chen and Pu Chen and Zhenyu Hu and Xingyu Li and Zhengxuan Jiang and Meng Cao and Tianfeng Long and Haozhe Wang and Mingzhang Wang and Yichen Zhang and Yiming Dai and Chenchen Zhang and Jiaying Wang and Zhiyong Wu and Shen Yan and Yujia Qin and Wenhao Huang and Zaiyuan Wang and Xiaolong Chang},
  year={2026},
  eprint={2606.11042},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2606.11042}
}