E-Commerce Bench: Evaluating LLM Agents on Long-Horizon Autonomous Business Operation
E-Commerce Bench is introduced, the first open-source benchmark that integrates multi-round counterpart negotiation and dynamic events into a year-long business operation, and it is found that no single model dominates.
Wei Fan, Xin-Jie Shen, Xu-Dong Guo et al.
· 2 citations