To determine whether sulfonylureas and exogenous insulin have different effects on insulin action, we studied eight patients with non-insulin-dependent diabetes mellitus before and after three months of treatment with tolazamide and exogenous semisynthetic human insulin, using a randomized crossover design. Therapy with tolazamide and therapy with insulin resulted in similar improvement of glycemic control, as measured by a decrease in mean glycosylated hemoglobin (+/- SEM) from 9.4 +/- 0.7 percent to 7.7 +/- 0.5 percent with tolazamide and to 7.1 +/- 0.2 percent with exogenous insulin (P less than 0.01 for both comparisons). Therapy with either tolazamide or exogenous insulin resulted in a similar lowering (P less than 0.05) of postabsorptive glucose-production rates (from 2.3 +/- 0.1 to 2.0 +/- 0.2 and 1.8 +/- 0.1 mg per kilogram of body weight per minute, respectively) but not to normal (1.5 +/- 0.1 mg per kilogram per minute). Both tolazamide and exogenous insulin increased (P less than 0.05) glucose utilization at supraphysiologic insulin concentrations (from 6.2 +/- 0.7 to 7.7 +/- 0.6 mg per kilogram per minute with tolazamide and to 7.8 +/- 0.6 mg per kilogram per minute with exogenous insulin) to nondiabetic rates (7.9 +/- 0.5 mg per kilogram per minute). Neither agent altered erythrocyte insulin binding at physiologic insulin concentrations. We conclude that treatment with sulfonylureas or exogenous insulin results in equivalent improvement in insulin action in patients with non-insulin-dependent diabetes mellitus. Therefore, the choice between these agents should be based on considerations other than their ability to ameliorate insulin resistance.