Objective: To identify predictors of ovarian response in women undergoing ovarian stimulation with corifollitropin alfa in a GnRH antagonist protocol and determine specific thresholds for the prediction of low and excessive responders. Design: Retrospective cohort study. Setting: University-based tertiary care center. Patient(s): Infertile women undergoing ovarian stimulation for in vitro fertilization/intracytoplasmic sperm injection. Intervention(s): Controlled ovarian hyperstimulation with corifollitropin alfa in a GnRH antagonist protocol. Main Outcome Measure(s): Relationship between ovarian reserve tests and ovarian response. Result(s): Antimüllerian hormone (AMH) and antral follicle count (AFC) were the only independent predictors for low and excessive ovarian response. In prediction of excessive response, the area under the receiver operating characteristic curve [AUC (95% CI)] for AMH was 0.890 (0.832-0.947) and 0.897 (0.829-0.964) for AFC. The optimal thresholds for identifying excessive responders were 3.52 ng/mL for AMH (sensitivity 89.5, specificity 83.8) and 16 for AFC (sensitivity 80.0, specificity 84.5). AMH and AFC also predicted low ovarian response: AUCs AMH 0.836 (0.783-0.889) and AFC 0.830 (0.767-0.894). The optimal thresholds for predicting low response were 1.37 ng/mL for AMH (sensitivity 74.1, specificity 77.5) and 8 for AFC (sensitivity 72.2, specificity 84.6). For both excessive and low ovarian responses, a logistic regression model combining the biomarkers was associated with improved discrimination. Conclusion(s): AMH and AFC are the best predictors for low and excessive response in women treated with corifollitropin alfa in an antagonist protocol. Using AMH and AFC to select suitable candidates for treatment with corifollitropin alfa may result in a safe and convenient stimulation. Copyright © 2013 American Society for Reproductive Medicine, Published by Elsevier Inc.