What causes host-use specificity in herbivorous insects? Population genetic models predict specialization when habitat preference can evolve and there is antagonistic pleiotropy at a performance-affecting locus. But empirically for herbivorous insects, host-use performance is governed by many genetic loci, and antagonistic pleiotropy seems to be rare. Here, we use individual-based quantitative genetic simulation models to investigate the role of pleiotropy in the evolution of sympatric host-use specialization when performance and preference are quantitative traits. We look first at pleiotropies affecting only host-use performance. We find that when the host environment changes slowly the evolution of host-use specialization requires levels of antagonistic pleiotropy much higher than what has been observed in nature. On the other hand, with rapid environmental change or pronounced asymmetries in productivity across host species, the evolution of host-use specialization readily occurs without pleiotropy. When pleiotropies affect preference as well as performance, even with slow environmental change and host species of equal productivity, we observe fluctuations in host-use breadth, with mean specificity increasing with the pervasiveness of antagonistic pleiotropy. So, our simulations show that pleiotropy is not necessary for specialization, although it can be sufficient, provided it is extensive or multifarious.