We comment on work by Ruszczynski et al., who extended the regression method of Cunico et al. (Fluid Phase Equilibria, 2014) and proposed certain Gibbs free energy models for binary liquid-liquid equilibria. We find that their regression method considers only necessary criteria for stability, and we derive constraints to be imposed on the thermodynamic parameters during regression such that the criteria become sufficient and the models thus become numerically easy to use.