In public health terms, leishmaniases are diseases of humans and dogs, whereas, in epidemiological terms, Leishmania spp. are considered to represent infections of a wide variety of animals, which represent the natural reservoirs of the various parasite species involved. Humans and dogs (which may he considered secondary or 'accidental' hosts in the leishmanial life-cycle) often exhibit severe clinical signs and symptoms when infected, whereas reservoir hosts generally show a few, minor or no signs. This situation makes the definition of a suitable laboratory model a difficult one, since the various experimental hosts may behave either like a reservoir or an accidental host.This review discusses the concept of animal models for leishmaniases and provides a critical evaluation of the most common experimental models and their respective advantages and disadvantages. In this state-of-the-art review, particular emphasis is given to the value of using mouse, hamster, cotton-rat, dog and primate models, especially in the context of testing potential anti-leishmanial vaccines.