 This paper introduces a new 3D point cloud classification benchmark data set containing over 4 billion manually labeled points. The authors also present submissions from the first participants using deep convolutional neural networks, CNNs. These submissions demonstrate significant improvement over existing state-of-the-art methods, suggesting that the data set can help unlock the full potential of deep learning methods for 3D labeling tasks. This article was authored by T. Hackel, N. Savanov, L. Ladiky and others.