Background: Women with polycystic ovarian syndrome (PCOS) are intrinsically insulin resistant and have a high risk of cardiovascular disease and type 2 diabetes. Weight loss improves risk factors, but the optimal diet composition is unknown. Low–glycemic index (low-GI) diets are recommended without evidence of their clinical effectiveness. Objective: We compared changes in insulin sensitivity and clinical outcomes after similar weight losses after consumption of a low-GI diet compared with a conventional healthy diet in women with PCOS. Design: We assigned overweight and obese premenopausal women with PCOS (n = 96) to consume either an ad libitum low-GI diet or a macronutrient-matched healthy diet and followed the women for 12 mo or until they achieved a 7% weight loss. We compared changes in whole-body insulin sensitivity, which we assessed using the insulin sensitivity index derived from the oral-glucose-tolerance test (ISIOGTT); glucose tolerance; body composition; plasma lipids; reproductive hormones; health-related quality of life; and menstrual cycle regularity. Results: The attrition rate was high in both groups (49%). Among completers, ISIOGTT improved more with the low-GI diet than with the conventional healthy diet (mean ± SEM: 2.2 ± 0.7 compared with 0.7 ± 0.6, respectively; P = 0.03). There was a significant diet-metformin interaction (P = 0.048), with greater improvement in ISIOGTT among women prescribed both metformin and the low-GI diet. Compared with women who consumed the conventional healthy diet, more women who consumed the low-GI diet showed improved menstrual cyclicity (95% compared with 63%, respectively; P = 0.03). Among the biochemical measures, only serum fibrinogen concentrations showed significant differences between diets (P < 0.05). Conclusion: To the best of our knowledge, this study provides the first objective evidence to justify the use of low-GI diets in the management of PCOS.