From the neutronic viewpoint, the optimization of BWR core designs is strongly related to the accurate determination of flux variations inside and around fuel assemblies. These fluctuations, which are mainly due to the high heterogeneity of the fuel and moderator regions, as additionally to the presence of cruciform absorber blades, have a direct impact on reactor safety and performance. Of particular importance is the pin power distribution, leading to the need of assessing the capabilities of design tools in a sufficiently rigorous manner. The basic configuration chosen for the code comparisons corresponds to a SVEA-96 fuel assembly under full-density water moderation conditions, with inserted hafnium absorber blades. The calculational schemes employed are the Monte Carlo code MCNPX2.5, in conjunction with various nuclear data libraries (ENDF/B-VI, JEF2.2, JEFF3.0, JENDL-3.2 and JENDL3.3), and the deterministic codes CASMO4 with JEF2.2, BOXER with JEF1.0 and HELIOS1.6 with ENDF/B-VI based libraries, respectively. The significant discrepancies observed in k, predictions (500pcm) are found to be mainly nuclear data related. On the other hand, data library effects have been found to be quite small for the prediction of pin-wise distributions of total fissions (Ftot), 238U captures (C8), as also of the C8/Ftot ratio. Significant differences in these reaction rate distributions (up to several percent) have, however, been observed between the Monte Carlo and deterministic calculations, particularly in the vicinity of the hafnium blades and in the gadolinium pins. Atomic Energy Society of Japan.