Until 1995, the etiology of ‘neonatal’ diabetes was totally unknown. In about a decade, mutations in 8 different genes (IPF1, EIF2AK3, GK, FOXP3, KCNJ11, ABCC8, PTF1A and GLIS3) have been discovered in patients with the permanent form of the disease, and 3 genetic abnormalities (defects in the paternally imprinted chromosomal region 6q24 and ‘mild’ activating mutations in KCNJ11 or ABCC8) have been detected in subjects with transient neonatal diabetes. Together with the advances in the understanding of the pathophysiology of this condition, clearly different from type 1 diabetes, also the temporal criterion by which one clinically defines a patient as being affected by neonatal diabetes has changed. In 1995, neonatal diabetes was defined as hyperglycemia that requires insulin treatment and occurs during the first month of life. In some patients with defects of KCNJ11, ABCC8 or EIF2AK3 genes however, diabetes can present at 6 months of age and beyond. It is now time to adopt a new definition in order to avoid the confusion originating by the misuse of the term ‘neonatal’. I would suggest monogenic diabetes of infancy, which includes both the permanent and the transient types, irrespectively of the mechanism of disease.