The importance of creative thinking was highlighted in the modern world by including it into the lists of 21st century skills and educational objectives of the national curricula. In order to make creativity assessment more widespread in schools and educational research, it is necessary to diminish the resource-intensity of the assessment for both educators and students. In modern psychometrics, automatic scoring of product outcomes is one of the urgent questions of creativity assessment. Automatic scoring increases the potential of scaling up since it does not depend on working hours of human raters and eliminates the bias due to experts’ subjectivity. Automatic scoring goes along with computer-based assessment. The aim of this research is to present a methodology of computer-based creative thinking assessment without human-based ratings. We have followed evidence-centered design and described a proficiency model, which is based on the cognitive approach and the concept of structural imagination, and the task models for computer-based assessment with automated scoring. In the empirical part, the factor structure of the tool was shown on a sample of 749 5th-grade students. We discussed the benefits of the proposed approach, such as an increase in ecological validity and a decrease in experts’ efforts. The practical significance is the decrease in workload in creativity assessment in classroom conditions using IT.