The objective. Assessment of the sensitivity and specificity of a method of calculating variable positions in HIV genome for detection of the duration of infection in a cohort of HIV-infected population of Russia. Patients and methods. The effectiveness of the method based on detection of HIV heterogenicity was assessed on 119 plasma specimens of HIV-infected patients with the known date of infection. The average duration of infection was 15 months, and the median - 8 months. In the studied sample, specimens of patients infected less than one year before amounted to 68% (81 of 119). The cohort consisted of 55 women (46%) and 64 men (54%). At the moment of blood testing for studying viral genome the patients' age varied from 0 to 79 years, the mean age was 31 years, median - 31 years. Results. As has been calculated for 119 examined specimens, the linear dependence between the duration of infection and the degree of variability of the sequence can be described by the formula y = 0.0012∗X + 0.0021 (R2 = 0.52). This equation permitted to calculate the variability threshold of 0.33%, which determined the duration of HIV infection as 12 months. Based on the obtained data, we calculated the sensitivity and specificity of the molecular method of detecting recent infection, which were 79.01 and 63.16%, respectively (comparable with previously published data). The duration of infection was calculated for each patient, then it was compared with the duration based on epidemiological data: for 42% of specimens the error in determining the duration of infection was less than 1 years, and for 92% of specimens - less than 3 years. Conclusion. This method might be successfully used in practice to assess the quality of screening programmes of detecting HIV infection performed in various regions and for various risk groups. Also, it might be used to confirm the presence of a recent (less than 1 year) infection.