In thts paper I argue that Kuhn's analysis of the role played by `exemplars' in normal science should be considered one of his most important contributions to the Philosophy of Science. The paper makes a detailed analysis of the relationship between the various elements of a 'disciplinary matrix´— empirical generalizations, models and exemplars — starting with Kuhn's views as exposed in various of his books and papers. Kuhn's analysis is investigated in the context of the discussions, in the 50s and 60s, about the notion of `model' in the so called `received view' of the structure of scientific theories. I refer also to the critics of logical empiricism at this time, that were concerned with analogical models and inspired by Campbell's studies on the dynamics of theories. I argue also that Polanyi's notion of 'tacit knowledge' and some of his discussions of the role played by analogies in mathematics have been an important influence on Kuhn. The paper indicates, furthermore, how Kuhn's intuitions concerning exemplars are being developped in the present, by plilosophers of science like Thagard, Paul Churchland and Giere, that adopt a `cognitivist´ stance towards scientific reasoning.