In a cross-sectional study, MSM aged ≥18 years were venue-sampled in New York City in 2008, interviewed, and tested for HIV using oral fluids. Participants who reported testing HIV negative at their last test in the previous 24 months were analyzed (n = 287 of 550 sampled). Those testing positive at the interview were defined as recently infected. HIV incidence was estimated using person-time at-risk methods and correlates of recent infection using proportional hazards regression. Thirty-two (11.1%) were recently infected. HIV incidence was 5.67/100 person-years at-risk. Independent correlates included: study recruitment in parks vs. bars, and in other venues vs. bars; black vs. non-black race/ethnicity; and reporting a last sex partner with a positive or unknown vs. negative HIV status. When assay-based methods are not feasible, cross-sectional HIV test results and self-reported HIV testing history and risk factor data can be used to estimate HIV incidence and the correlates of recent infection.