Eleven 233U solution critical assemblies spanning an H/233U ratio range of 40 to 2000 and an unreflected metal 233U assembly have been calculated with ENDF/B-IV and Hansen-Roach cross sections. Results from these calculations are compared with the experimental results and with each other. We observed an increasing disagreement between calculations with ENDF/B and Hansen-Roach data with decreasing H/233U ratio, indicative of large differences in their intermediate energy cross sections. The Hansen-Roach cross sections appeared to give reasonably good agreement with experiments over the whole range, whereas the ENDF/B calculations yielded high values for keff on assemblies of low moderation. We conclude that serious problems exist in the ENDF/B-IV representation of the 233U cross sections in the intermediate energy range and that further evaluation of this nuclide is warranted. In addition, we recommend that an experimental program be undertaken to obtain 233U criticality data at low H/233U ratios for verification of generalized criticality safety guidelines.