Acute erythroid leukemia (AEL) is characterized by a distinct morphology, mutational spectrum, lack of preclinical models, and poor prognosis. Here, using multiplexed genome editing of mouse hematopoietic stem and progenitor cells and transplant assays, we developed preclinical models of AEL and non-erythroid acute leukemia and describe the central role of mutational cooperativity in determining leukemia lineage. Different combination of mutations in Trp53, Bcor, Dnmt3a, Rb1, and Nfix resulted in the development of leukemia with an erythroid phenotype, accompanied by the acquisition of alterations in signaling and transcription factor genes that recapitulate human AEL by cross-species genomic analysis. Clonal expansion during tumor evolution was driven by mutational cooccurrence, with clones harboring a higher number of founder and secondary lesions (eg, mutations in signaling genes) showing greater evolutionary fitness. Mouse and human AEL exhibited deregulation of genes regulating erythroid development, notably Gata1, Klf1, and Nfe2, driven by the interaction of mutations of the epigenetic modifiers Dnmt3a and Tet2 that perturbed methylation and thus expression of lineage-specific transcription factors. The established mouse leukemias were used as a platform for drug screening. Drug sensitivity was associated with the leukemia genotype, with the poly (ADP-ribose) polymerase inhibitor talazoparib and the demethylating agent decitabine efficacious in Trp53/Bcor-mutant AEL, CDK7/9 inhibitors in Trp53/Bcor/Dnmt3a-mutant AEL, and gemcitabine and bromodomain inhibitors in NUP98-KDM5A leukemia. In conclusion, combinatorial genome editing has shown the interplay of founding and secondary genetic alterations in phenotype and clonal evolution, epigenetic regulation of lineage-specific transcription factors, and therapeutic tractability in erythroid leukemogenesis.