Background: We address the research questions: (i) what proportion of new HIV infections is transmitted from people who are (a) undiagnosed, (b) in primary HIV infection (PHI), (c) on antiretroviral therapy?; and (ii) what is the expected epidemiological impact of (a) increasing the proportion of newly acquired HIV infections receiving early treatment, and (b) increasing HIV testing rates? Methods: We used a mathematical model to simulate HIV transmission in the population of men who have sex with men (MSM) in Australia. We calibrated the model using established biological and clinical data and a wide range of Australian MSM epidemiological and behavioural data sources. Results: We estimate that ~19% of all new HIV infections are transmitted from the ~3% of Australian HIV-infected MSM who are in PHI; ~31% of new HIV infections are estimated to be transmitted from the ~9% of MSM with undiagnosed HIV. We estimate that the average number of infections caused per HIV-infected MSM through the duration of PHI is ~0.14–0.28. Conclusions: The epidemiological impact of increasing treatment in PHI would be modest due to insufficient detection of newly-infected individuals. In contrast, increases in HIV testing rates could have substantial epidemiological consequences. The benefit of testing will also increase over time. Promoting increases in the coverage and frequency of testing for HIV could be a highly-effective public health intervention, but the population-level impact of interventions based on promoting early treatment of patients diagnosed in PHI is likely to be small. Treating PHI requires further evaluation of its long-term effects on HIV-infected individuals.