Tailoring defense responses to different attackers is important for plant performance. Plants can use secondary metabolites with dual functions in resistance and defense signaling to mount herbivore-specific responses. To date, the specificity and evolution of this mechanism are unclear. Here, we studied the functional architecture, specificity, and genetic basis of defense regulation by benzoxazinoids in cereals. We document that DIMBOA-Glc induces callose as an aphid resistance factor in wheat. O-methylation of DIMBOA-Glc to HDMBOA-Glc increases plant resistance to caterpillars but reduces callose inducibility and resistance to aphids. DIMBOA-Glc induces callose in wheat and maize, but not in Arabidopsis, while the glucosinolate 4MO-I3M does the opposite. We identify a wheat O-methyltransferase (TaBX10) that is induced by caterpillar feeding and converts DIMBOA-Glc to HDMBOA-Glc in vitro. While the core pathway of benzoxazinoid biosynthesis is conserved between wheat and maize, the wheat genome does not contain close homologs of the maize DIMBOA-Glc O-methyltransferase genes, and TaBx10 is only distantly related. Thus, the functional architecture of herbivore-specific defense regulation is similar in maize and wheat, but the regulating biosynthetic genes likely evolved separately. This study shows how two different cereal species independently achieved herbivore-specific defense activation by regulating secondary metabolite production.