CooperBench: Why Coding Agents Cannot be Your Teammates Yet
CooperBench:为何编码代理还不能成为你的队友
机构 * Stanford University(斯坦福大学) ; SAP Labs US(SAP美国实验室)
AI总结 CooperBench通过大规模协作编码任务测试,发现AI代理在团队协作中表现不佳,揭示了沟通障碍、承诺偏离和期望错误等关键问题,呼吁发展社交智能。
Comments https://cooperbench.com First two authors contribute equally. The 3th - 6th authors contribute equally