Reproduce “EnterpriseBench: Benchmarking LLM Agents on Enterprise-Level Strategic Reasoning and Decision-Making” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Min Yang, Yichen Pan, Jinghua Piao, Dandan Song, Yongshun Gong, Yong Li
EnterpriseBench evaluates enterprise QA and interactive decisions through consulting interviews, Beer Game ordering, and project planning. Agents ask for hidden facts, manage delayed consequences, or set plans. Beer Game's lowest-cost method changes from Discussion at 1.94 × 10⁴ with DeepSeek-V3 to GEPA at 3.09 × 10⁴ with GPT-4.1. These results come from controlled text-based environments and do not establish readiness.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.