The aim of the study was to investigate students’ views of model evaluation through the lens of personal epistemology. We developed an integrated analytical framework by combining a developmental framework, including absolutist, multiplist, and evaluatist, with a multi-dimensional framework, including limits of knowing, certainty of knowing, and criteria of knowing. Furthermore, we examined the potential influence of the question contexts and the students’ grade levels. A total of 188 secondary school students were surveyed. Students answered two sets of model evaluation questions based on two scientific contexts. After reading the information about the two models, the students had to choose from three epistemic assumptions and then provide written justifications explaining their choice of assumptions. Quantitative and qualitative analyses were conducted for the multiple-choice questions and the written responses. In both contexts there were higher percentages of 11th-grade students choosing the evaluatist assumptions than the eighth-grade students. For students choosing multiplist and evaluatist assumptions, the 11th-grade students were more likely than the eighth-grade students to think in terms of pragmatic and evidential criteria as the criteria of knowing. Different contexts of the questions evoked different views of model evaluation particularly regarding the limits of knowing. Four additional categories of epistemic levels also emerged from the data. This study provides a new framework for understanding students’ thinking about model evaluation. Implications and suggestions for future research are provided.