The primary issue regarding the proliferation of radioactive materials is their possible ill-intentioned use. Depending on the material, it could enable the construction of dirty bombs or even nuclear devices. Several detection systems have been engineered to help control the transport of these materials and to provide efficient detection capabilities. Compton cameras have been used in fields such as medical or astronomical imaging for nearly 40 years. The existing research on the Compton-camera concept is unfortunately not applicable to these applications: the energies and distances of interest are very different. We have designed a new way to simulate a two-plane Compton camera for nuclear nonproliferation applications using the MCNP-PoliMi code. The simulations include accurate background models and detector properties such as time and energy resolutions, and pulse-generation time. Energy spectra can be obtained for both planes, along with the back-projection images. In this work, we present a study on the sensitivity of various large-scale Compton-camera configurations using this simulation tool. The Compton-camera materials investigated are PVT and CaF2 for the scatter plane, and NaI and LaBr3 for the absorption plane. The planes optimized in previous work; the voxels of 2 inch × 2 inch were used throughout this work. Results are presented for the weakest detectable source (1.4 mCi for CaF 2/NaI) at a 100-m standoff in a 60-s measurement for the various Compton-camera configurations.