3D models readily available on the Internet allow to perform visual localization, i.e., camera pose (position and orientation) estimation, right away without the time-consuming scene capturing and model building steps. Yet, it also comes with challenges as the available 3D models are often imperfect reflections of reality. This benchmark was created to help the community with the research on this underexplored source of data by aggregating the 3D meshes of multiple levels of detail and fidelity, corresponding image datasets and providing the means of evaluation and comparison.