Insulin molecules of size much greater than natural insulin have been synthesized and studied with the intention of widening the therapeutic window between adequate glycemic control and hypoglycemia as compared with conventional insulins. MK-1092 is a synthetic insulin dimer with favorable properties demonstrated in preclinical studies. Here, we report the results of the first-in-human, randomized, double-blind, active-control, single ascending dose trial of MK-1092, conducted in healthy adults, adults with type 1 diabetes (T1D), and adults with type 2 diabetes (T2D). MK-1092 was well tolerated in all study populations, and no dose-related adverse events were identified across the evaluated dose range (4-64 nmol/kg). Circulating concentrations of MK-1092 were approximately dose-proportional. Maximum glucose infusion rate (GIR) and 24-hour time-weighted average GIR were evaluated under euglycemic clamp conditions. These pharmacodynamic measurements were approximately dose-proportional in all study populations; at similar doses, the GIR parameters were lower in adults with T2D than in healthy adults or adults with T1D, likely due to the influence of insulin resistance. At doses ≥ 16 nmol/kg, MK-1092 had similar or greater effects than glargine 3 nmol/kg (0.5 units/kg) on increasing GIR in each study population and on suppressing free fatty acids and ketone generation in adults with T1D. MK-1092 did not prevent a subsequent high dose of lispro from increasing the GIR in healthy adults. Additional studies in adults with T1D and T2D are needed to further evaluate the safety, tolerability, and efficacy profile of MK-1092 and its potential for differentiation from more conventional insulins.