Learning on graphs is one of the most active research topics in machine learning (ML). Among the key challenges in this field, effectively learning long-range dependencies in graphs has been a particularly difficult problem. It has been observed that, in practice, the performance of many ML approaches, including various types of graph neural networks (GNNs), degrades significantly when the learning task involves long-range dependencies—that is, when the answer is determined by the presence of a certain path of significant length in the graph. This issue has been attributed to several phenomena, including, most prominently, oversmoothing, over-squashing, and vanishing gradient. A number of solutions have been proposed to mitigate these causes. However, evaluation of these solutions is complicated by the fact that existing benchmarks do not really test systems for their ability to learn tasks based on long-range dependencies in a transparent manner. In this paper, we design a synthetic benchmark that provably allows testing systems for this learning ability. We then evaluate state-of-the-art systems against it and conclude that none of them can claim that it can learn long-range dependencies well. We also observe that this weak performance cannot be attributed to any of the three causes, thus indicating that further investigation is necessary.