In Australia, HIV prevalence estimates among gay men have been mainly based on self-reported HIV status collected in annual behavioural surveys. We measured biological HIV prevalence among gay men in Melbourne, Australia, using a facility based sampling method. We calculated HIV prevalence and used logistic regression to assess correlates of a positive HIV test. A total of 639 gay men were recruited completed a survey and provided oral fluid for HIV testing from seven venues in 2008. The median age of the participants was 35 years (range 18?75 years). Overall biological HIV prevalence was 9.5 (95 CI 7.5?12.0 ) compared with 6.3 (95 CI 4.5?8.4 ) for self-reported HIV positive status. We found a significant discrepancy between test detected biological and self-report HIV status in our study, with 19 men (31.1 ) unaware of their HIV infection. These results highlight the importance of repeatable biological estimates to inform and evaluate HIV prevention strategies.