This study evaluates two significantly different models of a Westinghouse 414 reactor system using the TRAC-PF1/MOD1 computer code for a small-break loss-of-coolant accident (SBLOCA). A coarse threedimensional model of the reactor vessel is developed. In the coarse model, three of the four reactor coolant loops are combined into one loop. A detailed three-dimensional model of the reactor vessel is also developed. In the detailed model, each of the four coolant loops is modeled separately. Both models are run to steady-state convergence until the calculated system parameters are in good agreement. In addition, the steady-state results of both models closely match operational parameters given in the final safety analysis report. From the self-consistent steady-state conditions, a 60-s transient calculation is performed with each model. The transient simulates a 4-in. SBLOCA. The overall results of code predictions for the two models closely agree, and the vessel global parameters for the two models are also in good agreement. However, the computer times for the two calculations are significantly different. The detailed model provides additional information that is unavailable with the less detailed model, such as temperature and void fraction distributions throughout different regions of the vessel. During the 60-s transient, the upper head in the detailed model shows extensive voiding. The upper head in the coarse model also shows voiding; however, the extent and exact location of the voiding are not available in the coarse model. During this transient, the core region does not show extensive voiding; however, the detailed model shows some localized boiling. The results indicate that the coarse model is sufficient for 4-in. SBLOCA studies. The computer time associated with TRAC-PF1/MOD1 calculation of the extremely detailed model is ∼100 times longer than the coarse model.