Radial distributions of the total fission rate and the 238U-capture-to-total-fission (C8/Ftot) ratio were measured in SVEA-96+ and SVEA-96 Optima2 assemblies during the LWR-PROTEUS program. Fission rates predicted using MCNPX with JEFF-3.1 cross sections underestimated the measured values in the gadolinium-poisoned pins of the SVEA-96 Optima2 assembly; similarly, C8/Ftot ratios were overestimated in some gadolinium-poisoned pins of the SVEA-96+ assembly. A considerable effort was invested at the Paul Scherrer Institut to explain the discrepancies in gadolinium pins, without success. Recently, gadolinium cross sections were measured at the Rensselaer Polytechnic Institute by Leinweber et al. and differed significantly from current library values. ENDF/B-VII.0 gadolinium cross sections have currently been modified to include the new measurements, and these data have been processed with NJOY to yield files usable by MCNPX. Fission rates in the gadolinium-poisoned fuel pins of the SVEA-96 Optima2 pins were increased by 1.4 to 2.0% using the newly produced cross sections, yielding to a better agreement with the experimental values. Predicted C8/Ftot ratios were decreased on average by 1.7% in both clustered and unclustered groups of gadolinium-poisoned fuel pins of the SVEA-96+ assembly correcting the overpredictions previously reported in the clustered gadolinium pins. Earlier reported discrepancies observed in PROTEUS integral experiments, between measured and calculated reaction rates in the gadolinium-poisoned pins, might thus be due to inaccurate gadolinium cross sections. The PROTEUS results support the new thermal and epithermal gadolinium data measured by Leinweber et al.