Recently, a new theoretically based departure function for multi-fluid mixture models has been proposed, which allows for the combination of the multi-fluid mixture model with a model for the excess Gibbs energy [A. Jäger, I.H. Bell, C. Breitkopf, Fluid Phase Equilibria 469 (2018) 56–69]. In the aforementioned article, it is demonstrated that the combination of a multi-fluid mixture model with UNIFAC yields comparable or better results for phase equilibrium calculations than using standard mixing rules, i.e., linear mixing rules or Lorentz-Berthelot combining rules, for the binary mixtures of the components ethanol, ethane, carbon dioxide, propene, and benzene. In this work, this multi-fluid mixture model is combined with the predictive excess Gibbs energy model COSMO-SAC for the first time. Results for all binary mixtures of the components ethanol, ethane, carbon dioxide, propene, and benzene are compared to experimental data as well as to results of the multi-fluid mixture model with standard mixing rules. The combination of COSMO-SAC and the multi-fluid mixture model yields comparable or better results than the multi-fluid mixture model with standard mixing rules for the binary mixtures considered. One of the main advantages of multi-fluid mixture models in general is that any thermal equation of state can be used to model a component in the mixture. Multiparameter equations of state, the Soave-Redlich-Kwong equation of state, and the Peng-Robinson equation of state are employed for modeling the components in the mixtures. It is demonstrated that the mixture model also yields good results if instead of multiparameter equations of state cubic equations of state are used.