In this work, a turbulent gas-liquid stirred vessel is investigated by a fully predictive computational method, based on the Two-Fluid Model equations in the Reynolds Averaged Navier-Stokes formulation. The typical geometrical characteristics and operating conditions of aerated industrial fermenters are selected, which pose specific challenges due to the complex distribution of the gas phase and the cavity formation on the rear of the flat blades of the Rushton turbines. The results show that successful predictions of the gas cavities and the subsequent power consumption reduction in different operating conditions are obtained. The differences between the impellers positioned at different axial heights are assessed both in terms of gassed to ungassed power ratio and of gas flow rate treated by each turbine, showing that very a different behavior is obtained on the lowest impeller, with respect to the upper ones, thus providing a tool to quantify the individual impeller performances. The spatial distribution of the gas phase dispersed in the system and the resulting local mass transfer coefficient distribution are also discussed, observing a local decrease of the kLa due to the cavity formation. The presented model may contribute to the characterization of gas distribution, power requirements and local kLa in industrial fermenters, advancing the state of the art of Computational Fluid Dynamics tools in the context of fermentation intensification.