Preclinical studies of prostate cancer (CaP) have employed a genetically engineered mouse model, since there is no naturally occurring CaP in rodents. We have previously reported a new knock-in mouse adenocarcinoma prostate (KIMAP) model. In this study, we demonstrate that the new model possesses a tumor architecture of heterogeneity and multifocality similar to that of human CaP, by utilizing a new compound scoring system to compare with the PSP94 (approved gene symbol Msmb) gene-directed transgenic mouse CaP model (TGMAP). KIMAP mice showed a balanced distribution of tumor extent, which penetrated the prostate gland. Comparative studies on cDNA microarrays demonstrated that KIMAP tumors were upregulated with higher contents of immunoresponse genes, whereas PSP-TGMAP tumors had neuroendocrine (NE) differentiation. The majority of KIMAP mice did not progress to NE CaP, which was observed only at a very late stage and a low frequency. Several tumor marker genes characteristic of human CaP were uniquely identified in KIMAP tumors, including hepsin, maspin, Nkx3.1, CD10 and PSP94 (similar to PSA), etc. The differences between these two CaP models are attributed to the introduction of a single endogenous knock-in mutation. Due to the similarities between human CaP tumors and the PSP-KIMAP tumors, this preclinical model may supplement the current transgenic models to study CaP more accurately.