This work proposes a model to explain concrete anisotropic ablation by corium during a molten core–concrete interaction (MCCI). As a result of recent MCCI prototypic material experiments, core-concrete interaction (CCI) tests, and VULCANO tests, one observes that concrete ablation behavior consistently depends on the concrete materials used in the experiments. Specifically, tests with limestone-common-sand (LCS) concrete yielded isotropic concrete ablation, i.e., equal axial and radial concrete erosion. This is in comparison to anisotropic ablation in tests with siliceous (SIL) concrete, where radial ablation was much larger than axial ablation. This was an unexpected result because prior results of many MCCI simulant experiments indicated that nearly isotropic ablation was expected in prototypic material experiments regardless of concrete type. A new phenomenological model is proposed in this work based on a hypothesis that unifies the result of both previous simulant and prototypic material experiments, i.e., heat transfer area enhancement and delayed gas release caused by the presence of unmelted solid aggregate material that enters the molten pool. This model offers a logical and phenomenological explanation concerning anisotropic ablation as well as the capability to simulate anisotropic ablation. This model is implemented into the CORQUENCH code as part of this work. Comparisons of these simulation results obtained with this new model to the CCI experiments for cases with SIL concrete and anisotropic ablation show better agreement with the test data than the existing model.