AgentBench
收藏arXiv2025-09-30 收录
下载链接:
https://github.com/thudm/agentbench
下载链接
链接失效反馈官方服务:
资源简介:
该数据集提供了一系列基准测试,用于评估代理在不同任务中的表现。此外,评估遵循由AgentBench建立的评估Lumos的常规方法,任务是评估代理的性能。
This dataset provides a suite of benchmark tests to evaluate the performance of AI agents across various tasks. Furthermore, the evaluation follows the standard methodology for assessing Lumos established by AgentBench, which targets evaluating the performance of AI agents.



