Preclinical models that faithfully recapitulate the genomic and histopathological features of cancer are critical for the development of new treatments. The most commonly used models are two-dimensional cell lines established from primary tumors or fluids. While these have provided some important insights into cancer biology, these cell models have significant limitations. In order to address some of these limitations, spheroids, tumor-derived organoids and microfluidic chips have more recently been used to investigate the role of the three-dimensional environment. Efforts have also been made to develop animal models, including genetically modified mice and patient-derived xenografts. We will highlight strengths and weaknesses of the available in vitro and in vivo models.