This paper presents an overview of the two previous editions of the benchmark conferences (2009 and 2011) with a focus on experimental trials results and on the comparison of the accuracy of the different FEM codes in predicting critical process outputs. Indeed benchmark experiments are designed with the aim to focus the attention on a particular phenomenon and to check the ability of FEM codes of predicting it. Profile lengths, process load, die deflections, die, and profile temperatures which were accurately recorded during the extrusion strokes were used as benchmarking parameters for FEM comparison. Moreover, computational times, simulation set-up times, as well as information on the required hardware were compared for the proposed cases study. A detailed discussion for all output parameters is described in order to illustrate the evolution of the potentiality and limits of each code used in the last three years.