We have performed a benchmark to evaluate the relative success of using small-angle X-ray scattering (SAXS) data as constraints (hereafter termed SAXSconstrain) in the RosettaDock protocol (hereafter termed RosettaDockSAXS). For this purpose, we have chosen 38 protein complex structures, calculated the theoretical SAXS data for the protein complexes using the program CRYSOL, and then used the SAXS data as constraints. We further considered a few examples where crystal structures and experimental SAXS data are available. SAXSconstrain were added to the protocol in the initial, low-resolution docking step, allowing fast rejection of complexes that violate the shape restraints imposed by the SAXS data. Our results indicate that the implementation of SAXSconstrain in general reduces the sampling space of possible protein-protein complexes significantly and can indeed increase the probability of finding near-native protein complexes. The methodology used is based on rigid-body docking and works for cases where no or minor conformational changes occur upon binding of the docking partner. In a wider perspective, the strength of RosettaDockSAXS lies in the combination of low-resolution structural information on protein complexes in solution from SAXS experiments with protein-protein interaction energies obtained from RosettaDock, which will allow the prediction of unknown three-dimensional atomic structures of protein-protein complexes.