Development of a nuclear thermal propulsion (NTP) engine is pursued by government and industry. Current deployment regulations require that experimental efforts are heavily complemented and guided by modeling and simulation. Challenges arise when directly simulating the physics of such systems at steady-state operation and especially during transients due to strongly coupled local and global effects. Reduced-order methods are commonly used as a practical means for obtaining full-core solutions, however these methods must be verified against higher-fidelity solvers. Therefore, a Python package, ANTHMM, has been developed that is intended for general usability and allows users to externally couple Monte Carlo neutronics solvers and commercial CFD software. In its current state, ANTHMM is optimized for the Serpent Monte Carlo code and Ansys Fluent CFD. Using published data from the XE-Prime NTP engine, a CFD model has been generated, validated, and used as a reference model for a fully coupled reduced-order solver. The CFD model and reduced order solver show excellent agreement for the hot-condition test data in three different regions of the core, which were recorded during the XE-Prime experiment. The neutronic behavior predicted by the fully coupled reduced order solver is very similar to that predicted by the coupled CFD model. Finally, it is shown that informed prediction from a reduced order solver of first-iteration temperatures and densities used in the MC simulation can reduce the coupled simulation from several iterations to a single predictor-corrector step.