This article describes a proposal for a framework to evaluate and compare enterprise models. It suggests three major categories for grouping the model evaluation criteria: syntactic, semantic and pragmatic analysis. The paper draws on a wide literature to present a large selection of criteria and to operationalise their measurement by means of several possible metrics. As an empirical validation test, a selection of metrics for eight of the criteria has been calculated for fifteen large enterprise models. Their interpretation supports the usefulness and validity of the overall framework. Various attempts at deriving a composite overall quality score are discussed, but there is less confidence in the validity of this approach.