Reinforced concrete (RC) structures with tri-dimensional asymmetries tend to exhibit torsional effects that are of great concern in the field of earthquake engineering, in particular at large ductility levels where they become more relevant [MAN09]. In nuclear facilities, this issue assumes particular relevance considering that these structures are designed to respond essentially in the elastic range with a controlled level of deformations and accelerations when subjected to strong ground motions. Within the previous framework, the research project SMART 2013 (‘Seismic design and best-estimate Methods Assessment for Reinforced concrete buildings subjected to Torsion and non-linear effects’) was conducted in order to improve the knowledge on the seismic response of irregular RC structures (experimental test) and to provide reference data for modelling developments and validation (benchmark). Past blind prediction contests showed that, by making use of appropriate modelling options, the seismic response of structures can be predicted with satisfactory accuracy [SOU14]. Nonetheless, they have also evidenced a significant dispersion of predictions between the different participating teams, as demonstrated by the results obtained in the past blind prediction contest [NEE10]. After a short description of the SMART 2013 mock-up and corresponding loading, the present paper describes the properties of the numerical model used by one the participating teams and the comparison of its results with those obtained from the experiments.