This paper presents a review of near-field mixing zone models and compares their performance with common outfall diffuser examples in the Great Lakes. The results of initial dilutions for three types of multiport diffusers, based on the Ontario Ministry of Environment (MOE) guidelines and recommendations, are compared with those calculated by using the U.S. Environmental Protection Agency (EPA) models. These models include five integral-type models (UPLUME, UOUTPLM, UDKHDEN, UMERGE, and ULINE) and a "length scale" type expert system (CORMIX2). Results based on the first four EPA integral models give higher initial dilutions when compared to results based on CORMIX2 and the MOE guidelines, ULINE gives unrealistically low dilutions. Excluding UPLUME, the initial dilutions given by the other EPA models increase with increasing ambient current. Alternating diffusers generally give lower initial dilutions than the staged and unidirectional diffusers, while unidirectional diffusers produce the highest of the three. Results from the computation tests based on the mixing models can aid in more reasonable and economical diffuser designs that still meet the requirement of initial dilution criteria. While all the models selected for this study have limitations, CORMIX2 is preferred for most Great Lakes applications. It can be applied to all four types of multiport diffusers and can handle different types of ambient stratification, bottom and shore attachment, wake effects and dead zones, plume trapping and far-field behaviour. CORMIX2 predictions compare well with laboratory data and very limited field data. Given the complexity of the problem, more field studies should be performed for further validation of the models. Key words: mixing zone, multiport diffusers, initial dilution.