The aim of the study was to investigate progress in reasoning capacity and knowledge base appraisal in a longitudinal analysis of data from summative evaluation throughout a medical problem-based learning curriculum. The scores in multidisciplinary discussion of a clinical case and multiple choice questionnaires (MCQs) were studied longitudinally for 213 students from years 2 to 5. The capacity of core knowledge delimitation was calculated as the difference between the levels of average ascertainment degrees given for correct and incorrect answers at MCQ. For both multidisciplinary discussion of a clinical case evaluation and self-estimation of core knowledge, the capacity increases throughout the curriculum. The reasoning capacity assessed through multidisciplinary discussion of a clinical case is positively correlated with MCQ scores and the capacity to discriminate the mastered core knowledge. In conclusion, this study indicates that self-estimation of core knowledge is associated with an increase in reasoning performance through a well-organised knowledge base. Since that ability is related to success or failure, it is suggested that student awareness about delimitation of mastered core knowledge is considered as part of learning.