Pulmonary nodules are a common clinical problem. The goals of their evaluation are to expedite the diagnosis and treatment of patients with malignant nodules and to minimize testing in patients with benign nodules. The approach to lung nodule evaluation is directed by the probability that the nodule is malignant. Estimation of the probability of malignancy can be performed subjectively with intuition and clinical experience or by using validated probability models that combine patient clinical characteristics with nodule imaging features to estimate a probability of malignancy. The accuracy and the generalizability of probability models depend on the clinical profile and the prevalence of malignancy in the population in which they were derived. In this article, we review available validated models to estimate the probability of malignancy in patients with pulmonary nodules and outline how they were derived, their limitations, and how they compare with each other and physician judgment. We conclude with a brief discussion of advances in probability models.