Beyond Static Tools: Test-Time Tool Evolution for Scientific Reasoning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909987779903488 |
|---|---|
| author | Lu, Jiaxuan Kong, Ziyu Wang, Yemin Fu, Rong Wan, Haiyuan Yang, Cheng Lou, Wenjie Sun, Haoran Wang, Lilong Jiang, Yankai Wang, Xiaosong Sun, Xiao Zhou, Dongzhan |
| author_facet | Lu, Jiaxuan Kong, Ziyu Wang, Yemin Fu, Rong Wan, Haiyuan Yang, Cheng Lou, Wenjie Sun, Haoran Wang, Lilong Jiang, Yankai Wang, Xiaosong Sun, Xiao Zhou, Dongzhan |
| contents | The central challenge of AI for Science is not reasoning alone, but the ability to create computational methods in an open-ended scientific world. Existing LLM-based agents rely on static, pre-defined tool libraries, a paradigm that fundamentally fails in scientific domains where tools are sparse, heterogeneous, and intrinsically incomplete. In this paper, we propose Test-Time Tool Evolution (TTE), a new paradigm that enables agents to synthesize, verify, and evolve executable tools during inference. By transforming tools from fixed resources into problem-driven artifacts, TTE overcomes the rigidity and long-tail limitations of static tool libraries. To facilitate rigorous evaluation, we introduce SciEvo, a benchmark comprising 1,590 scientific reasoning tasks supported by 925 automatically evolved tools. Extensive experiments show that TTE achieves state-of-the-art performance in both accuracy and tool efficiency, while enabling effective cross-domain adaptation of computational tools. The code and benchmark have been released at https://github.com/lujiaxuan0520/Test-Time-Tool-Evol. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_07641 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Beyond Static Tools: Test-Time Tool Evolution for Scientific Reasoning Lu, Jiaxuan Kong, Ziyu Wang, Yemin Fu, Rong Wan, Haiyuan Yang, Cheng Lou, Wenjie Sun, Haoran Wang, Lilong Jiang, Yankai Wang, Xiaosong Sun, Xiao Zhou, Dongzhan Artificial Intelligence Computation and Language Multiagent Systems The central challenge of AI for Science is not reasoning alone, but the ability to create computational methods in an open-ended scientific world. Existing LLM-based agents rely on static, pre-defined tool libraries, a paradigm that fundamentally fails in scientific domains where tools are sparse, heterogeneous, and intrinsically incomplete. In this paper, we propose Test-Time Tool Evolution (TTE), a new paradigm that enables agents to synthesize, verify, and evolve executable tools during inference. By transforming tools from fixed resources into problem-driven artifacts, TTE overcomes the rigidity and long-tail limitations of static tool libraries. To facilitate rigorous evaluation, we introduce SciEvo, a benchmark comprising 1,590 scientific reasoning tasks supported by 925 automatically evolved tools. Extensive experiments show that TTE achieves state-of-the-art performance in both accuracy and tool efficiency, while enabling effective cross-domain adaptation of computational tools. The code and benchmark have been released at https://github.com/lujiaxuan0520/Test-Time-Tool-Evol. |
| title | Beyond Static Tools: Test-Time Tool Evolution for Scientific Reasoning |
| topic | Artificial Intelligence Computation and Language Multiagent Systems |
| url | https://arxiv.org/abs/2601.07641 |