Metastatic breast cancer (mBC) is deadly, and its molecular drivers are largely unknown. Treatment is limited to systemic cytotoxic chemotherapies and tumors often become refractory. The most frequently mutated gene in MBC is TP53. The TP53R248W hotspot missense mutation is associated with poor prognosis. We generated a somatic mouse model of mammary epithelial specific Trp53R245W expression. Primary tumors were highly metastatic, reflecting the human molecular subtypes luminal A, luminal B, HER2, and TNBC. Transcriptomic profiling revealed three intrinsic subtypes: stem-cell like (SCL), well-differentiated metabolically active (WDMA), and immunosuppressed (IS). SCL tumors activate ribosome biosynthesis and E2F signaling, amplifying Met, Birc3, Yap1 and deleting Nf1, Pik3r1, and Rad17. WDMA tumors activate cytochrome P450 enzymes, estrogen signaling and branched chain amino acid degradation, with mutations activating Pi3k/Akt/mTOR signaling. IS tumors activate immune suppression, have high mutation burden, and frequently mutate Traf7 and delete Cdkn2a. This is the most comprehensive transcriptomic and genomic profiling of mutant p53-driven breast tumors, elucidating potential therapeutic targets.