Neuro-symbolic artificial intelligence is a novel area of AI research which\nseeks to combine traditional rules-based AI approaches with modern deep\nlearning techniques. Neuro-symbolic models have already demonstrated the\ncapability to outperform state-of-the-art deep learning models in domains such\nas image and video reasoning. They have also been shown to obtain high accuracy\nwith significantly less training data than traditional models. Due to the\nrecency of the field's emergence and relative sparsity of published results,\nthe performance characteristics of these models are not well understood. In\nthis paper, we describe and analyze the performance characteristics of three\nrecent neuro-symbolic models. We find that symbolic models have less potential\nparallelism than traditional neural models due to complex control flow and\nlow-operational-intensity operations, such as scalar multiplication and tensor\naddition. However, the neural aspect of computation dominates the symbolic part\nin cases where they are clearly separable. We also find that data movement\nposes a potential bottleneck, as it does in many ML workloads.\n