Model-driven development tools built on industry standard platforms, such as the EclipseModeling Framework (EMF), heavily utilize model queries in model transformation, well-formedness constraint validation and domain-specific model execution. As these queries are executed rather frequently in interactive modeling applications, they have a significant impact on runtime performance and end user experience. However, due to their complexity, these queries can be time consuming to implement and optimize on a case-by-case basis. Consequently, there is a need for a model query framework that combines an easy-to-use and concise declarative query formalism with high runtime performance. In this paper, we propose a declarative EMF model query framework using the graph pattern formalism as the query specification language. These graph patterns describe the arrangement and properties of model elements that correspond to, e.g. a well-formedness constraint, or an application context of a model transformation rule. For improved runtime performance, we employ incremental pattern matching techniques: matches of patterns are stored and incrementally maintained upon model manipulation. As a result, query operations can be executed instantly, independently of the complexity of the constraint and the size of the model. We demonstrate our approach in an industrial (AUTOSAR) model validation context and compare it against other solutions.