[en] As part of an independent assessment study to establish code capabilities, comparisons between the RELAP4/MOD6, UPDATE 4 pre-test prediction and experimental data are presented for Semiscale Mod-3 Reflood Test S-07-4. Where deficiencies are indicated in the code-calculated representation of the experiment, methods of code and input selection criteria improvement are developed and appropriate recommendations made. Additional studies include the effect of updating test conditions from planned to actual and of sensitivity to downcomer nodalization