A new benchmark has been introduced for evaluating automated domain modeling approaches. The benchmark combines two existing datasets, allowing researchers to compare their methods across different levels of complexity and scale. It provides a natural language description and corresponding ground truth domain model, enabling the evaluation of generated models against reference models using a metric. This benchmark is intended to address the lack of standardized benchmarks in this area, facilitating future research on automated domain modeling. --- Why it matters: This matters because it enables researchers to compare their approaches more effectively, which can lead to better understanding and improvement of automated domain modeling techniques. Source: https://arxiv.org/abs/2608.15255