Despite the fact that scientific models and works of fiction are the products of the same tools, i.e. idealization or approximation, models differ from works of fiction because the idealization practices in science act as guiding instruments that lead to new knowledge. This epistemic function is characteristic of scientific models and it is a necessary condition for the success of a model, whether theory-driven or phenomenological. It is not a necessary condition for works of fiction because, when these works represent, they represent only the general features of their target thus shaping or improving our intuitions about those features of the respective targets; they do not necessarily lead to knowledge concerning the specific features of the target. Hence, if we treat scientific models as works of fiction in order to understand what models are and how they function, we run the risk of overlooking this distinct kind of epistemic function that models display.