Breast cancer is a heterogeneous disease with many subtypes, and the association between these subtypes and exposure to environmental factors such as radiation remains controversial. Although the rat is used widely for research into human breast cancer, the heterogeneity and subtype definitions are unclear. Here, we leveraged an archive of rat mammary cancer samples and gene expression microarray data to classify tumors and examine their association with exposures. Eighty-four mammary cancer and 12 normal mammary tissue samples were obtained from previous experiments in which rats were exposed to different types of radiation, chemical carcinogens, and diets. Tumors were then subjected to immunohistochemical (IHC) analysis of conventional biomarkers, as well as gene expression profiling; they were then classified by three approaches based on IHC results, the PAM50 classifier algorithm, and unsupervised clustering of gene expression profiles. IHC identified four subtypes (luminal A-like, luminal B-like 1, luminal B-like 2, and triple-negative), while PAM50 identified six (luminal A, luminal B, basal-like, HER2-enriched, normal-like, and claudin-low). Unsupervised clustering divided the tumors into three large, statistically significant, groups (named "luminal A", "luminal B", and "non-luminal" clusters). The results of the three approaches were significantly associated with each other. Exposure to radiation and chemical carcinogens during post-pubertal development was significantly associated with an increased risk of developing luminal A tumors, whereas exposure to a high corn-oil diet was associated with a higher likelihood of luminal B tumors. Rat mammary cancer subtypes resemble those in humans and are related to environmental factors.