Low-cost generic antiretroviral drugs are available in resource-limited settings for treatment of HIV infections. However, few bioequivalence data in specific populations in which these generics are likely to be used are available. We conducted a randomized crossover bioequivalence study of generic and brand name formulations of nevirapine, zidovudine, and lamivudine in HIV-negative Indian women using US Food and Drug Administration (FDA) criteria. Subjects took single doses of all formulations separated by a 14-day washout period. Plasma concentrations were measured over 96 hours during each study period. Average bioequivalence was determined using natural log-transformed maximum concentration (C(max)) and area-under-the-concentration-time curve (AUC) mean ratio data. Fifteen Indian women were enrolled. The 90% confidence intervals for nevirapine (14 subjects) and lamivudine (15 subjects) C(max), AUC from 0 to the last measurable time point (AUC(0-t)), and AUC from 0 to infinity (AUC(0-infinity)) mean ratios and zidovudine (15 subjects) AUC(0-t) and AUC(0-infinity) mean ratios were all within 0.80 to 1.25. However, the 90% confidence interval for zidovudine C(max) mean ratio was 0.70 to 1.46. Generic and brand name nevirapine and lamivudine met FDA average bioequivalence criteria. Lack of average bioequivalence for zidovudine was found for C(max) but is not expected to be clinically significant, because the total AUC values were similar between formulations.