AgentBaton：智能体任务中的高效模型协作

AI 摘要

该论文研究长程智能体任务中强模型与高效模型的协作方式，系统比较了模型交接的方向与上下文传递方法。实验发现，高效模型到强模型的交接通常不如仅用强模型，而强模型到高效模型在合适阶段交接可大幅降低成本并保持甚至提升准确率。基于此提出轻量检测器 AgentBaton，在 11 组模型对和两个基准上平均降低 API 成本 65.9%，准确率最高提升 8.8 个百分点。

正文节选

AgentBaton: Efficient Model Collaboration in Agentic Tasks Abstract Solving long-horizon agentic tasks requires language agents to interact with their environments over multiple steps, and reliable performance often comes from the expensive capable models available through cloud providers. Meanwhile, alternative efficient models with inferior performance but a much lower cost are emerging, including lower-tier cloud models and smaller open-weights models deployable on local hardware. This raises

发布时间：2026-10-28 12:00

抓取时间：2026-10-28 12:07

来源机构：arXiv