Adaptation is an important element on the climate change policy agenda. Integrated assessment models, which are key tools to assess climate change policies, have begun to address adaptation, either by including it implicitly in damage cost estimates, or by making it an explicit control variable. We analyze how modelers have chosen to describe adaptation within an integrated framework, and suggest many ways they could improve the treatment of adaptation by considering more of its bottom-up characteristics. Until this happens, we suggest, models may be too optimistic about the net benefits adaptation can provide, and therefore may underestimate the amount of mitigation they judge to be socially optimal. Under some conditions, better modeling of adaptation costs and benefits could have important implications for defining mitigation targets.