Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2605.26086 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910255681634304 |
|---|---|
| author | Lin, Yusong Liang, Xinyuan Wang, Haiyang Gu, Qipeng Cheng, Siqi Chen, Jiangui Wu, Shuzhe Pan, Feiyang Fan, Lue Zhao, Sanyuan Tu, Dandan |
| author_facet | Lin, Yusong Liang, Xinyuan Wang, Haiyang Gu, Qipeng Cheng, Siqi Chen, Jiangui Wu, Shuzhe Pan, Feiyang Fan, Lue Zhao, Sanyuan Tu, Dandan |
| contents | Large language model agents are increasingly envisioned as always-on personal assistants with access to anything relevant in the user's digital world. Yet current systems operate over only narrow slices of that world, limiting context-sensitive reasoning and effective assistance. Existing benchmarks similarly provide only partial user state and therefore fail to capture performance in such a broad, always-on setting. To address this gap, we introduce Claw-Anything, a benchmark that expands agent context along three dimensions: long-horizon activity histories, interdependent backend services, and integrated GUI and CLI interaction across multiple devices. To instantiate this setting, we simulate months of user activity through multi-round event injection, producing complex world states and realistic noise, including irrelevant events and conflicting signals. Agents must reason over rich contextual environments while remaining robust to such noise. This expanded scope also enables the evaluation of proactive assistance, requiring agents to anticipate user needs and deliver timely recommendations. Experiments show that GPT-5.5 achieves only 34.5% pass@1, substantially below prior benchmarks, underscoring a gap between current agent capabilities and the demands of always-on personal assistance. Alongside the benchmark, we release an automated data-generation pipeline that yields 2,000 training environments and improves the base model by 23.7%, demonstrating its utility of scalable data infrastructure. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_26086 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Claw-Anything: Benchmarking Always-On Personal Assistants with Broader Access to User's Digital World Lin, Yusong Liang, Xinyuan Wang, Haiyang Gu, Qipeng Cheng, Siqi Chen, Jiangui Wu, Shuzhe Pan, Feiyang Fan, Lue Zhao, Sanyuan Tu, Dandan Artificial Intelligence Large language model agents are increasingly envisioned as always-on personal assistants with access to anything relevant in the user's digital world. Yet current systems operate over only narrow slices of that world, limiting context-sensitive reasoning and effective assistance. Existing benchmarks similarly provide only partial user state and therefore fail to capture performance in such a broad, always-on setting. To address this gap, we introduce Claw-Anything, a benchmark that expands agent context along three dimensions: long-horizon activity histories, interdependent backend services, and integrated GUI and CLI interaction across multiple devices. To instantiate this setting, we simulate months of user activity through multi-round event injection, producing complex world states and realistic noise, including irrelevant events and conflicting signals. Agents must reason over rich contextual environments while remaining robust to such noise. This expanded scope also enables the evaluation of proactive assistance, requiring agents to anticipate user needs and deliver timely recommendations. Experiments show that GPT-5.5 achieves only 34.5% pass@1, substantially below prior benchmarks, underscoring a gap between current agent capabilities and the demands of always-on personal assistance. Alongside the benchmark, we release an automated data-generation pipeline that yields 2,000 training environments and improves the base model by 23.7%, demonstrating its utility of scalable data infrastructure. |
| title | Claw-Anything: Benchmarking Always-On Personal Assistants with Broader Access to User's Digital World |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2605.26086 |