Survival of patients with advanced solid tumors has not significantly improved over the past 30 years. Although molecularly targeted anticancer drugs offer promise, few drugs make it through the end of the Food and Drug Administration approval process. Animal models that more closely resemble human carcinogenesis may bridge the gap between preclinical success and benefits for patients. We discuss pros and cons of several mouse models, including genetically engineered mice that each represent different aspects of human cancer, and the screening of targeted drugs in these models.