Model verification, or the process of ensuring that the prescribed equations are properly solved, is a necessary step in code development. Careful, quantitative verification guides users when selecting grid resolution and time step and gives confidence to code developers that existing code is properly instituted. This work introduces the RadBelt radiation belt model, a new, open-source version of the Dynamic Radiation Environment Assimilation Model (DREAM) and uses the Method of Manufactured Solutions (MMS) to quantitatively verify it. Order of convergence is investigated for a plethora of code configurations and source terms. The ability to apply many different diffusion coefficients, including time constant and time varying, is thoroughly investigated. The model passes all of the tests, demonstrating correct implementation of the numerical solver. The importance of DLL and source term dynamics on the selection of time step and grid size is also explored. Finally, an alternative method to apply the source term is examined to illustrate additional considerations required when non-linear source terms are used.