To gain insight into HIV transmission we estimated the proportion of those recently infected. We examined data from HIV-positive patients and a random 10% sample of HIV-negative patients tested at Voluntary Counseling and Testing sites in Poland in 2006. Archived samples from positive patients were tested by three assays to differentiate recent from long-standing infection. Using logistic regression, we examined the association of recent infection (at least one assay) with age, sex, HIV exposure category, and the interval between self-reported HIV exposure and previous HIV test. Of 13,511 tests, 154 (1.1%) were HIV positive, representing 19.7% (n=783) of new diagnoses in Poland in 2006. Demographic and behavioral data were linked for 95, of whom 45 (47%) were recently infected and 1,001 were HIV negative. New diagnoses were more likely to be injectors (17% vs. 2%), men who have sex with men (MSM) (37% vs. 12%), and less frequent condom users (7.8% vs. 14% always) compared to HIV negatives. The median number of partners during the past 12 months was one and two among positives and negatives, but was higher among MSM-four and three, respectively. Ever injectors were less likely to be recently infected (adjusted OR=0.15, 95%CI=0.03-0.73). Having two or more sexual partners in the past 12 months was an independent predictor of recent infection (4.01, 1.4-11.49). We found no evidence that age or sex predicted recent infection. These data reinforce health education campaigns for safe sex messages, especially among MSM. They also suggest, albeit based on a subset of new diagnoses, that interventions should not be limited to selected age/sex groups.