代理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管理训练数据知之甚少。现有的开放工作（例如 SWE-Smith, SERA, 和 Nemotron-Terminal）通常针对单个基准,，从而留下了如何训练泛化于不同代理任务的模型的问题。 OpenThoughts-Agent (OT-Agent) 项目通过用于训练代理模型的完全开放的数据管理管道来解决这一差距。我们进行了 100 多个受控消融实验，系统地研究管道, 的每个阶段，从而深入了解任务来源和多样性的重要性。然后，我们从管道中收集了 10 万个示例的训练集，并在此数据集, 上微调 Qwen3-32B，在七个代理基准上产生的平均准确度为 44.8%，比现有最强的开放数据代理模型 (Nemotron-Terminal-32B, 40.9%) 提高了 3.9 个百分点。此外,我们的训练数据表现出强大的扩展特性,在计算控制的比较中在每个训练集大小上都优于替代开放数据集。我们在此 http URL 公开发布我们的训练集, 数据管道, 实验数据, 和模型，以支持未来关于代理模型训练的开放研究。
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 this http URL to support future open research on agentic model training.
科目:人工智能(cs.AI)
Subjects: Artificial Intelligence (cs.AI)