**摘要**
Agentic语言模型极大地扩展了人工智能的应用，但公众对如何为具有广泛能力的代理人策划培训数据知之甚少。现有的开放性工作，如SWE-Smith、SERA和Nemotron-Terminal ，通常针对单个基准，留下了如何训练跨越不同代理任务的模型的问题。OpenThoughts-Agent （ OT-Agent ）项目
👤 作者: Negin Raoof, Richard Zhuang, Marianna Nezhurina, Etash Guha, Atula Tejaswi, Ryan Marten, Charlie F. Ruan, Tyler Griggs, Alexander Glenn Shaw, Hritik Bansal, E. Kelly Buchanan, Artem Gazizov, Reinhard Heckel, Chinmay Hegde, Sankalp Jajee, Daanish Khazi, Emmanouil Koukoumidis, Xiangyi Li, Hange Liu, Shlok Natarajan, 我叫Harsh Raj, Nicholas Roberts, Ethan Shen, Nishad Singhi, Michael Siu, Ashima Suvarna, Hanwen Xing, Patrick Yubeaton, Robert Zhang, Leon Liangyu Chen, Xiaokun Chen, Steven Dillmann, Saadia Gabriel, Xunyi Jiang, Anurag Kashyap, 李博轩, Yein Park, Minh Pham, Sujay Sanghavi, Lin Shi, Ke Sun, Yixin Wang, Zhiwei Xu, Erica Zhang, Siyan Zhao, Wanjia Zhao, Jenia Jitsev, Alex Dimakis, Benjamin Feuer, 路德维希·施密特

---
🔗 **[OpenThoughts-Agent ：代理模型的数据配方](https://arxiv.org/abs/2606.24855v1)**

> OpenThoughts-Agent: Data Recipes for Agentic Models
🏷️ 来源: ArXiv cs.AI
⏱️ 2026-06-24 14:02