For proving the precision achievable today in predicting the neutron fluence for reactor pressure vessels, the OECD/NEA initiated a new international benchmark exercise as a blind test, where the measured results are not known in advance. To validate two- and three-dimensional radiation transport codes, high precision measurements at the VENUS critical facility of SCK.CEN, Mol in Belgium were used representing the pressure vessel internals of a Westinghouse three loop reactor. About 20 independent calculations were contributed world-wide for the two-dimensional benchmark VENUS-1. The most challenging task was the validation of the newest versions of three-dimensional transport codes like TORT, PENTRAN and MCNP. 14 calculations were contributed for the complex 3D-benchmark VENUS-3 and compared with about 400 local measurements. In conclusion, the results of the three-dimensional benchmark VENUS-3 were in general much closer to the experimental values than for the two-dimensional benchmark VENUS-1, which shows inherent deficiencies of two-dimensional transport calculations that suffer on approximations like diffusion based axial buckling or improper averaging of axial sources and shielding layers.