Historically, HIV vaccines specifically designed to raise cellular immunity resulted in protection from disease progression but not infection when tested in monkeys challenged with a single high virus exposure. An alternative approach, more analogous to human sexual exposures, is to repetitively challenge immunized monkeys with a much lower dose of virus until systemic infection is documented. Using these conditions to mimic human sexual transmission, we found that a multi-protein DNA/MVA HIV-1 vaccine is indeed capable of protecting rhesus monkeys against systemic infection when repeatedly challenged with a highly heterologous immunodeficiency virus (SHIV). Furthermore, this repetitive challenge approach allowed us to calculate per-exposure probability of infection, an observed vaccine efficacy of 64%, and undertake a systematic analysis for correlates of protection based on exposures needed to achieve infection. Therefore, improved non-human primate models for vaccine efficacy can provide novel insight and perhaps renew expectations for positive outcomes of human HIV clinical trials.