paper-with-me

Papers

OpenThoughts-Agent: Data Recipes for Agentic Models

2026-06-23 · 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, 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, Boxuan Li, 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, Ludwig Schmidt arxiv

Agentic language models dramatically expand the applications of AI yet little is publicly known about how to curate training data for broadly capable agents. Existing open efforts such as SWE-Smith, SERA, and Nemotron-Terminal typically target a single benchmark, leaving open the question of how to train models that generalize across diverse agentic tasks. The OpenThoughts-Agent (OT-Agent) project addresses this gap with a fully open data curation pipeline for training agentic models. We conduct more than 100 controlled ablation experiments to systematically investigate each stage of the pipeline, yielding insights on the importance of task sources and diversity. We then assemble a training set of 100K examples from our pipeline and fine-tune Qwen3-32B on this dataset, which yields an average accuracy of 44.8% across seven agentic benchmarks and a 3.9 percentage point improvement over the strongest existing open data agentic model (Nemotron-Terminal-32B, 40.9%). Moreover, our training data exhibits strong scaling properties, outperforming alternative open datasets at every training set size in compute-controlled comparisons. We publicly release our training sets, data pipeline, experimental data, and models at openthoughts.ai to support future open research on agentic model training.

📄 PDF Abstract BibTeX arXiv:2606.24855

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

OpenThoughts: Data Recipes for Reasoning Models

2025-06-04 · Etash Guha, Ryan Marten, Sedrick Keh, Negin Raoof 외

Reasoning models have made rapid progress on many benchmarks involving math, code, and science. Yet, there are still many open questions about the best training recipes for reasoning since state-of-the-art models often r…

Math

Orchard: An Open-Source Agentic Modeling Framework

2026-05-14 · Baolin Peng, Wenlin Yao, Qianhui Wu, Hao Cheng 외 arxiv

Agentic modeling aims to transform LLMs into autonomous agents capable of solving complex tasks through planning, reasoning, tool use, and multi-turn interaction with external environments. We present Orchard, an open-so…

Demystifying Reinforcement Learning in Agentic Reasoning

2025-10-13 · Zhaochen Yu, Ling Yang, Jiaru Zou, Shuicheng Yan 외 arxiv

Recently, the emergence of agentic RL has showcased that RL could also effectively improve the agentic reasoning ability of LLMs, yet the key design principles and optimal practices remain unclear. In this work, we condu…

Reinforcement Learning

rStar2-Agent: Agentic Reasoning Technical Report

2025-08-28 · Ning Shang, Yifei Liu, Yi Zhu, Li Lyna Zhang 외 arxiv

We introduce rStar2-Agent, a 14B math reasoning model trained with agentic reinforcement learning to achieve frontier-level performance. Beyond current long CoT, the model demonstrates advanced cognitive behaviors, such …

Reinforcement Learning

DR-Venus: Towards Frontier Edge-Scale Deep Research Agents with Only 10K Open Data

2026-04-21 · Venus Team, Sunhao Dai, Yong Deng, Jinzhen Lin 외 arxiv

Edge-scale deep research agents based on small language models are attractive for real-world deployment due to their advantages in cost, latency, and privacy. In this work, we study how to train a strong small deep resea…

Reinforcement Learning