OpenThoughts-Agent：智能體模型的數據配方

摘要

代理语言模型大幅擴展了人工智慧的應用範圍，但關於如何為具廣泛能力的代理模型精選訓練資料，目前公開資訊仍相當有限。現有開放專案如 SWE-Smith、SERA 及 Nemotron-Terminal 通常僅針對單一基準進行最佳化，對於如何訓練能泛化至多樣代理任務的模型，仍留下未解問題。OpenThoughts-Agent（OT-Agent）專案針對此缺口，提出一套完全開放的資料精選管線，用於訓練代理模型。我們進行了超過 100 次控制消融實驗，系統性地探討管線中各個階段，並從中獲得關於任務來源與多樣性重要性的洞見。接著，我們利用此管線組建了一個包含 10 萬筆範例的訓練集，並以該資料集微調 Qwen3-32B，結果在七個代理基準上取得平均 44.8% 的準確率，相較現有最強的開放資料代理模型（Nemotron-Terminal-32B，40.9%）提升了 3.9 個百分點。此外，我們的訓練資料展現了強烈的擴展特性：在計算資源受控的比較中，所有訓練集規模下皆優於其他開放資料集。我們已在 openthoughts.ai 公開釋出訓練集、資料管線、實驗資料及模型，以支持未來關於代理模型訓練的開放式研究。

English

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.