Understanding, exploring, and interacting with the world through models characterizes science in all its branches and at all levels of education. Model-based reasoning is central to science education and thus science assessment. Building on research in assessment, science education, and learning sciences, this report provides a set of desgin patterns to help assessment designers, researchers, and teachers create tasks for assessing aspects of model-based reasoning: Model Formation, Model Use, Model Elaboration, Model Articulation, Model Evaluation, Model Revision, and Model-Based Inquiry. Each design pattern lays out considerations concerning targeted knowledge and ways of capturing and evaluating students' work. The ideas are illustrated with examples from existing assessments and the research literature.