In this work, we present a new Geant4-based simulation modeling platform for complex nuclear medicine imaging scanners that uses computer-aided design (CAD) tools and parallelization on both CPU and GPU to accelerate the Monte Carlo simulation and reduce computational cost. The application combines the newly developed open-source GDML package and the new version of open-source Opticks code to create simulation models for complex structures. We demonstrate that the application is suitable for imaging scanners with complex structures, and we show that the computational cost is significantly decreased. Our results show the computational efficiency enhanced over two orders of magnitudes using both GPU and CPU. Our application has the potential to make the simulation of complex imaging systems more accessible and efficient for new users of nuclear Monte Carlo codes. Also, we plan to extend this work to include other modalities and to examine the performance of more complex structures.