A benchmark for evaluating LLM agents on stateful, multi-step planning and tool use across realistic enterprise workflows.

The dataset is hosted on HuggingFace and can be loaded using the datasets library. from datasets import load_dataset

ds = load_dataset("EnterpriseAgents/EnterpriseOpsGym", split="train")

To filter by domain:

calendar_tasks =… See the full description on the dataset page: https://huggingface.co/datasets/EnterpriseAgents/EnterpriseOpsGym.