Diabetes has increasingly been recognized as a heterogeneous disease, with clinical characteristics and outcomes risk varying across different phenotypes. Evidence on heterogeneity of gestational diabetes (GDM) is yet to be provided. To investigate the insulin physiology and pregnancy outcomes of GDM phenotypes characterized by fasting hyperglycemia or postload hyperglycemia. A total of 2050 women who underwent a 75-g oral glucose tolerance test were prospectively recruited and followed up until delivery. Women were categorized into normoglycemia (NGT, n = 936), isolated impaired fasting glucose (gestational-IFG, n = 378), and isolated impaired postload glucose tolerance (gestational-IGT, n = 736) groups. Fasting blood sample at mid-pregnancy were collected to measure C-peptide and insulin concentrations. Homeostasis model assessment (HOMA) and quantitative insulin sensitivity check index (QUICKI) were used to evaluate insulin physiology. Maternal and neonatal outcomes were recorded. Gestational-IFG had greater insulin resistance (HOMA-IR 3.11 vs. 2.25, QUICKI-CP 0.94 vs. 1.03, both P < 0.01), and gestational-IGT had worse β-cell function (C-peptide 2.00 vs. 2.26 ng/ml, P < 0.05) when compared to one another. Gestational-IFG was more strongly associated with excessive gestational weight gain (RR 1.62, 95% CI 1.18-2.23) and large-for-gestational-age infants (RR 1.45, 95% CI 1.03-2.03) than gestational-IGT. The risk for neonatal brain injury was increased in gestational-IGT (RR 2.03, 95% CI 1.04-4.09), but not in gestational-IFG (P = 0.439). Gestational-IGT showed a stronger association with the risk of preterm birth compared to gestational-IFG (RR 1.80, 95% CI 1.02-3.36). GDM exhibits distinct insulin physiology profiles. Pregnancy outcome varies between each phenotype. These findings provide evidence on risk stratification and diverse strategies for the treatment of GDM.