Over the past 20 years, a variety of models have been developed to simulate the bioconcentration of hydrophobic organic chemicals by fish. These models differ not only in the processes they address but also in the way a given process is described. Processes described by these models include chemical diffusion through the gill's interlamellar water, epithelium, and lamellar blood plasma; advective chemical transport to and from the gill by ventilation and perfusion, respectively; and internal chemical deposition by thermodynamic partitioning to lipid and other organic phases. This article reviews the construction and associated assumptions of 10 of the most widely cited fish bioconcentration models. These models are then compared with respect to their ability to predict observed uptake and elimination rates using a common database for those model parameters that they have in common. Statistical analyses of observed and predicted exchange rates reveal that rates predicted by these models can be calibrated almost equally well to observed data. This fact is independent of how well any given model is able to predict observed exchange rates without calibration. The importance of gill exchange models and how they might by improved are also discussed.