This is Part II of a two part paper on the development of a parameterized linear magnetic equivalent circuit (MEC) for radial flux magnetic gears with surface permanent magnets. Part I describes the MEC implementation. This section, Part II, evaluates the MEC model's accuracy by comparing its results against those produced by a nonlinear finite-element analysis (FEA) model. Simulation results demonstrate that the linearity approximation does not prevent the MEC from accurately matching the torque and air-gap flux densities predicted by a nonlinear FEA model for three diverse magnetic gear base designs. The impacts of the MEC discretization parameters introduced in Part I are also investigated using the same base designs, and guidelines for those settings are developed. Additionally, single design parameter sweeps illustrate the MEC's ability to track these changes over most practical design ranges and reveal where the MEC's accuracy degrades due to the linearity approximation. Finally, the results of a 46,656 case parametric optimization study demonstrate the MEC's ability to match the nonlinear FEA model's torque predictions within a few percent over a wide range of designs while achieving average simulation speeds 44 to 271 times faster than those of the FEA model.