Climate models are not just physics translated into computer code. They are powerful actors influencing and influenced by humans. Thus modelers need to learn and modeling courses need to teach not only the techniques of numerical discretisation and the physical understanding of the climate system, but also the underlying motivations, the uncertainties and the societal embededness of the modeling approach. Following a design-based research approach, this study develops a course at Bachelor level that aims to teach students such interdisciplinary perspectives. With a reflective open-ended exercise, we elicit students' learning process through challenging climate modeling topics. We find that the students learn to appreciate the complexity of climate models and the intricacies of scientific practice itself, highlighting for example the role of values in science. The exercise reveals few misconceptions and no major hurdles in the students' learning that may have been expected from the interdisciplinary nature of the material. We thus conclude that the course is a practice-proven approach to teaching the physical basis of climate modeling as well as its critical reflection. Together with the openly shared material, it supplies an inspiration and practical template for lecturers to include more interdisciplinary content and reflection into their modeling courses.