OBJECTIVE To evaluate two different criteria, one or two cut-off values, of oral glucose tolerance test with 75g of glucose for the diagnosis of gestational diabetes mellitus. METHODS A cross-sectional study involving 120 records of pregnant women who received prenatal care at the service of a Brazilian university was carried out. Bivariate analysis of obstetric and perinatal outcomes was performed using the chi-square test. RESULTS Considering criterion I, 12.5% of patients were diagnosed with gestational diabetes mellitus. Patients were 3.57 times more likely to have a large fetus for the gestational age at birth (p=0.038). Using criterion II, gestational diabetes mellitus was diagnosed in 5.8% of patients, macrosomia was 7.73 times more likely to be found in the presence of gestational diabetes mellitus (p=0.004), and a large fetus for the gestational age at birth was 8.17 times more likely (p=0.004). CONCLUSIONS There was a difference in the prevalence of gestational diabetes mellitus between the two criteria analyzed. The new criterion proposed increased prevalence.