Cochlear models may be effectively used to predict the generation of OAEs, by simulating the cochlear response to stimuli of different waveform, frequency and (for nonlinear models) levels. As the cochlea is a nonlinear compressive system, OAE suppression by tones and noise is a common phenomenon, which must be correctly predicted in order to correctly interpret the experiments and to design acquisition systems, dedicated to localize the sources or to elicit a residual. Indeed, accurate predictions of the measurable properties (amplitude, phase, I/O functions, etc.) of the OAE response may help both in the interpretation of experimental results, and in the design of experimental and diagnostic protocols. In some cases, the predictions are partly model independent, because they rely only on the validity of general properties such as the cochlear scaling symmetry; in other cases, the predictions are dependent on fine details of the models, which further justifies the continuous efforts dedicated to refining theoretical modeling.