When do models provide genuine understanding, and why does it matter?

In spite of a number of remarkable advances over the last few decades, the question of how scientific models provide explanations remains rather controversial. In the case of biology, this question is quite pressing, since according to many specialists we do not find genuine universal explanatory laws in biology, and yet, biologists constantly use models as explanatory devices. It is probably not possible to reduce all of the ways in which models explain to one unifying pattern. If there is a common trait in all of these different ways it is perhaps that they achieve a better understanding of phenomena. It is argued in this paper that the notion of "understanding" is not irremediably subjective. Various contextual but objectively justifiable criteria are suggested in order to establish when a feeling of understanding corresponds to a genuine understanding.