With the advances in molecular genetics, animal models of human diseases are becoming more numerous and more refined every year. Despite this, one must recognize that they generally do not faithfully and comprehensively mimic the homologous human disease. Faced with these imperfections, some geneticists believe that these models are of little value, while for others, on the contrary, they are important tools. We agree with this second statement, and in this review, we examine the reasons that may explain the observed differences and suggest means to circumvent or even exploit them. Our opinion is that animal models should be regarded more as tools capable of answering specific questions rather than mere replicas, at a smaller scale, of a given human disease. Far from disappointing they are probably called for a promising future.