Given the role of the CD4 T helper cells in the development of memory CTL precursors, it seems bene"cial to boost the CD4 T helper response in the context of vaccination against the human immunode"ciency virus (HIV). However, CD4 T cells are also the preferred targets of infection by HIV. Here, we address the question as to whether it is advantageous to stimulate the CD4 T helper cell response, as this will increase the pool of potential target cells of infection. To do so we formulated a mathematical model describing the interactions between virus-infected cells, susceptible cells, HIV-speci"c CD4 helper T cells, and CTL precursor (CTLp) and e!ector cells (CTLe). The e!ect of increased initial CD4 helper and CTLp numbers on the outcome of infection, as well as the e!ect on viral set point of increased CD4 T helper growth rate, CTL responsiveness and the rate at which CTLp and CTLe are produced were studied. We found that only when the virus has a low basic reproductive number does the number of CTLp and CD4 T helper cells at the moment of infection in#uence the outcome of infection. In this situation, high initial T helper and CTL numbers can switch the outcome from full-blown infection to virus control. However, this holds for virus with infectivity in a limited range, and current estimates of virus infectivity suggest that it is higher. In that case, only a vaccination protocol that increases CTL responsiveness, ideally in combination with the rate of production of CD4 T helper cells, may o!er a solution as it can reduce the viral set point considerably. If brought under a certain level, the viral population might be unable to replicate any further. However, changing these parameters of the immune response is only bene"cial when infection is controlled by CTL in the long term. When a CD4 lymphoproliferative response is mounted but the CTL response is not maintained, increasing the CD4 T helper growth rate is deleterious.