EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910152191377408 |
|---|---|
| author | Zhu, Xinyu Cai, Yuzhu Liu, Zexi Wang, Cheng Li, Fengyang Jin, Wenkai Liu, Wanxu Bing, Zehao Zheng, Bingyang Chai, Jingyi Tang, Shuo Ye, Rui Du, Yuwen Pang, Xianghe Du, Yaxin Miao, Tingjia Zhang, Yuzhi Liao, Ruoxue Ding, Zhaohan Zhang, Linfeng Wang, Yanfeng E, Weinan Chen, Siheng |
| author_facet | Zhu, Xinyu Cai, Yuzhu Liu, Zexi Wang, Cheng Li, Fengyang Jin, Wenkai Liu, Wanxu Bing, Zehao Zheng, Bingyang Chai, Jingyi Tang, Shuo Ye, Rui Du, Yuwen Pang, Xianghe Du, Yaxin Miao, Tingjia Zhang, Yuzhi Liao, Ruoxue Ding, Zhaohan Zhang, Linfeng Wang, Yanfeng E, Weinan Chen, Siheng |
| contents | The convergence of large language models and agents is catalyzing a new era of scientific discovery: Agentic Science. While the scientific method is inherently iterative, existing agent frameworks are predominantly static, narrowly scoped, and lack the capacity to learn from trial and error. To bridge this gap, we present EvoMaster, a foundational evolving agent framework engineered specifically for Agentic Science at Scale. Driven by the core principle of continuous self-evolution, EvoMaster empowers agents to iteratively refine hypotheses, self-critique, and progressively accumulate knowledge across experimental cycles, faithfully mirroring human scientific inquiry. Crucially, as a domain-agnostic base harness, EvoMaster is exceptionally easy to scale up -- enabling developers to build and deploy highly capable, self-evolving scientific agents for arbitrary disciplines in approximately 100 lines of code. Built upon EvoMaster, we incubated the SciMaster ecosystem across domains such as machine learning, physics, and general science. Evaluations on four authoritative benchmarks (Humanity's Last Exam, MLE-Bench Lite, BrowseComp, and FrontierScience) demonstrate that EvoMaster achieves state-of-the-art scores of 41.1%, 75.8%, 73.3%, and 53.3%, respectively. It comprehensively outperforms the general-purpose baseline OpenClaw with relative improvements ranging from +159% to +316%, robustly validating its efficacy and generality as the premier foundational framework for the next generation of autonomous scientific discovery. EvoMaster is available at https://github.com/sjtu-sai-agents/EvoMaster. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_17406 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale Zhu, Xinyu Cai, Yuzhu Liu, Zexi Wang, Cheng Li, Fengyang Jin, Wenkai Liu, Wanxu Bing, Zehao Zheng, Bingyang Chai, Jingyi Tang, Shuo Ye, Rui Du, Yuwen Pang, Xianghe Du, Yaxin Miao, Tingjia Zhang, Yuzhi Liao, Ruoxue Ding, Zhaohan Zhang, Linfeng Wang, Yanfeng E, Weinan Chen, Siheng Artificial Intelligence The convergence of large language models and agents is catalyzing a new era of scientific discovery: Agentic Science. While the scientific method is inherently iterative, existing agent frameworks are predominantly static, narrowly scoped, and lack the capacity to learn from trial and error. To bridge this gap, we present EvoMaster, a foundational evolving agent framework engineered specifically for Agentic Science at Scale. Driven by the core principle of continuous self-evolution, EvoMaster empowers agents to iteratively refine hypotheses, self-critique, and progressively accumulate knowledge across experimental cycles, faithfully mirroring human scientific inquiry. Crucially, as a domain-agnostic base harness, EvoMaster is exceptionally easy to scale up -- enabling developers to build and deploy highly capable, self-evolving scientific agents for arbitrary disciplines in approximately 100 lines of code. Built upon EvoMaster, we incubated the SciMaster ecosystem across domains such as machine learning, physics, and general science. Evaluations on four authoritative benchmarks (Humanity's Last Exam, MLE-Bench Lite, BrowseComp, and FrontierScience) demonstrate that EvoMaster achieves state-of-the-art scores of 41.1%, 75.8%, 73.3%, and 53.3%, respectively. It comprehensively outperforms the general-purpose baseline OpenClaw with relative improvements ranging from +159% to +316%, robustly validating its efficacy and generality as the premier foundational framework for the next generation of autonomous scientific discovery. EvoMaster is available at https://github.com/sjtu-sai-agents/EvoMaster. |
| title | EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2604.17406 |