The structures present in the upper plenum of a sodium cooled fast reactor are potentially subject to thermal striping. In fact there is the possibility of a significant temperature variation at the exit of two adjacent subassemblies since they are typically hydro-dynamically isolated in a SFR core. It has been demonstrated over the years that it is extremely difficult to produce generalized heuristic models for the prediction of thermal striping due to the inherent dependency of such phenomena upon the geometry and local flow conditions. Thus the development and validation of numerical tools such CFD to predict thermal striping is crucial for safety evaluation. Thermal striping is however very challenging for CFD due to the presence of buoyancy effects and turbulent mixing, which usually require a separate treatment to achieve satisfying results. Fundamental experiments targeted specifically to CFD validation are thus necessary to select the appropriate models and methodologies for the simulation offlows involving thermal striping. The present work is concerned with steady-state and unsteady CFD simulations of an experiment currently scheduled to be performed at Argonne National Laboratory. The test section will be instrumented with particle image velocimetry (PIV) capable of extracting instantaneous velocity distributions over any given cross section in the domain and with a high-speed thermal imaging camera that will provide real-time temperature distributions on the upper surface of the test section. When operational, data from the MAX experiment will be compared with LES and RANS simulations. The calculations presented in this work were performed in support of the experiment definition and development. In particular the focus will be on the apparent strong sensitivity of the numerical results upon the modeling of the jet inlets.