Transgenic and knockout mice have been proposed as substitutes for one of the standard 2-yr rodent assays. The advantages of using genetically engineered mouse models is that fewer mice are needed, the time to develop disease is greatly reduced, and the mice are predisposed to developing cancer by virtue of gain or loss of functions. The models currently being used have yielded a large amount of data and have proved to be informative for risk assessment; however, they are still far from ideal. In fact, they inherently do not reflect the complexity of mutation and carcinogenesis in humans. Recent advances in technology and the creation of new knockout mice may produce more useful and more sensitive models. This review covers two recent advances in technology--inducible and regulatable gene expression and targeted genetic modifications in the genome--that will allow us to make better models. I also discuss new gene deletion and transgenic mouse models and their potential impact on risk-assessment assays. These models are presented in the context of four basic components or events that occur in the multistep process leading to cancer: maintenance of gene expression patterns, genome stability and DNA repair, cell-cell communication and signaling, and cell-cycle regulation. Finally, surrogate markers and utility in risk assessment are also discussed. This review is meant to stimulate further discussion in the field and to generate excitement about working toward the next generation of risk-assessment models.