Summary | Inelastic modeling of entire reinforced concrete (RC) buildings under seismic excitation is a complex problem that influences directly the predicted seismic performance. Modeling assumptions and conventions adopted become more important in existing RC frame response predictions, due to these structures’ structural characteristics and non conforming detailing. The problem is investigated for a typical existing five-story RC frame which has been designed for moderate seismicity according to the past generation of Greek seismic codes. Different plane frame finite element models are formulated adopting state of the art as well as state of the practice analysis codes and finite element formulations. The seismic performance of each model is estimated, following both a conventional static pushover as well as nonlinear timehistory analyses under different levels of seismic intensity. The models range from the simple yet widely adopted in practice concentrated plasticity elements with axial-flexural strength interaction only, to the more complex distributed damage stiffness or flexibility-based fiber elements accounting or not for joint deformations. The results of the analyses are compared at the global and primarily the local damage prediction levels, to reveal substantial discrepancies and scatter in key performance Response Indices introduced in a Performance Based (re)Design approach by the model limitations, which are often ignored. It is concluded that, in addition to standardization of the criteria and procedures of evaluation, the analytical model for evaluating these Response Indices should also be well defined to avoid error and conflict.