This paper presents an assessment conducted at the Paul Scherrer Institut (PSI) of two Monte Carlo-based codes, namely Monteburns and Serpent, for LWR depletion calculations. First, a validation of predicted spent fuel compositions is presented on the basis of a selected Swiss spent fuel sample from the LWR-PROTEUS Phase II program. Second, an additional assessment is performed for an OECD/NEA burnup credit benchmark which is also related to isotopic predictions. In both cases, equivalent pin-cell models are applied and the results are presented in terms of relative isotopic concentrations between measured/benchmark and predicted values. Additional comparisons are also made to previously published PSI results using the MCNPX/CINDER and CASMO-4E codes. The results show that all codes provide an overall good agreement with experimental data although a few specific trends are observed for some nuclides which are pointed out and briefly discussed.