ClawBench: Can AI Agents Complete Everyday Online Tasks?
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908949793472512 |
|---|---|
| author | Zhang, Yuxuan Wang, Yubo Zhu, Yipeng Du, Penghui Miao, Junwen Lu, Xuan Xu, Wendong Hao, Yunzhuo Cai, Songcheng Wang, Xiaochen Zhang, Huaisong Wu, Xian Lu, Yi Lei, Minyi Zou, Kai Yin, Huifeng Nie, Ping Chen, Liang Jiang, Dongfu Chen, Wenhu Allen, Kelsey R. |
| author_facet | Zhang, Yuxuan Wang, Yubo Zhu, Yipeng Du, Penghui Miao, Junwen Lu, Xuan Xu, Wendong Hao, Yunzhuo Cai, Songcheng Wang, Xiaochen Zhang, Huaisong Wu, Xian Lu, Yi Lei, Minyi Zou, Kai Yin, Huifeng Nie, Ping Chen, Liang Jiang, Dongfu Chen, Wenhu Allen, Kelsey R. |
| contents | AI agents may be able to automate your inbox, but can they automate other routine aspects of your life? Everyday online tasks offer a realistic yet unsolved testbed for evaluating the next generation of AI agents. To this end, we introduce ClawBench, an evaluation framework of 153 simple tasks that people need to accomplish regularly in their lives and work, spanning 144 live platforms across 15 categories, from completing purchases and booking appointments to submitting job applications. These tasks require demanding capabilities beyond existing benchmarks, such as obtaining relevant information from user-provided documents, navigating multi-step workflows across diverse platforms, and write-heavy operations like filling in many detailed forms correctly. Unlike existing benchmarks that evaluate agents in offline sandboxes with static pages, ClawBench operates on production websites, preserving the full complexity, dynamic nature, and challenges of real-world web interaction. A lightweight interception layer captures and blocks only the final submission request, ensuring safe evaluation without real-world side effects. Our evaluations of 7 frontier models show that both proprietary and open-source models can complete only a small portion of these tasks. For example, Claude Sonnet 4.6 achieves only 33.3%. Progress on ClawBench brings us closer to AI agents that can function as reliable general-purpose assistants. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_08523 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ClawBench: Can AI Agents Complete Everyday Online Tasks? Zhang, Yuxuan Wang, Yubo Zhu, Yipeng Du, Penghui Miao, Junwen Lu, Xuan Xu, Wendong Hao, Yunzhuo Cai, Songcheng Wang, Xiaochen Zhang, Huaisong Wu, Xian Lu, Yi Lei, Minyi Zou, Kai Yin, Huifeng Nie, Ping Chen, Liang Jiang, Dongfu Chen, Wenhu Allen, Kelsey R. Computation and Language Artificial Intelligence AI agents may be able to automate your inbox, but can they automate other routine aspects of your life? Everyday online tasks offer a realistic yet unsolved testbed for evaluating the next generation of AI agents. To this end, we introduce ClawBench, an evaluation framework of 153 simple tasks that people need to accomplish regularly in their lives and work, spanning 144 live platforms across 15 categories, from completing purchases and booking appointments to submitting job applications. These tasks require demanding capabilities beyond existing benchmarks, such as obtaining relevant information from user-provided documents, navigating multi-step workflows across diverse platforms, and write-heavy operations like filling in many detailed forms correctly. Unlike existing benchmarks that evaluate agents in offline sandboxes with static pages, ClawBench operates on production websites, preserving the full complexity, dynamic nature, and challenges of real-world web interaction. A lightweight interception layer captures and blocks only the final submission request, ensuring safe evaluation without real-world side effects. Our evaluations of 7 frontier models show that both proprietary and open-source models can complete only a small portion of these tasks. For example, Claude Sonnet 4.6 achieves only 33.3%. Progress on ClawBench brings us closer to AI agents that can function as reliable general-purpose assistants. |
| title | ClawBench: Can AI Agents Complete Everyday Online Tasks? |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2604.08523 |