Harmonization of biomarkers is important for the comparability of laboratory results as it allows the definition of universal reference values and clinical decision limits. In diabetology, immunoassays are widely used to determine HbA1c, C-peptide, insulin, and autoantibodies to beta cell proteins, which are essential biomarkers for the diagnosis and classification of diabetes mellitus. Furthermore, as large clinical studies have identified HbA1c as a predictor for the development of diabetic complications, HbA1c has evolved as the general treatment target. For decades, the use of non-harmonized assays caused confusion. After the standardization of HbA1c, the worldwide comparability improved and increased the confidence in this laboratory biomarker. Insulin and C-peptide are not only valuable biomarkers to assess beta-cell function, but may also be used to evaluate insulin resistance, a metabolic feature of type 2 diabetes often occurring before its manifestation. Long-lasting efforts led to substantial improvements in the harmonization process of C-peptide assays, but harmonization of insulin assays is still ongoing. Therefore, C-peptide is now sometimes used as a surrogate biomarker for insulin. Furthermore, autoantibodies against beta cell components are important biomarkers for the accurate differentiation of type 1, type 2, and other special types of diabetes. Owing to the heterogeneity of these autoantibodies against beta cell proteins, harmonization is very difficult to achieve. International efforts are in progress to harmonize the current assays, as the presence of autoantibodies against beta cell proteins predicts the development of type 1 diabetes in early life. In conclusion, clinical studies linking diagnosis, classification, prediction, and treatment to laboratory values of the respective biomarkers need to be harmonized to avoid misdiagnosis and incorrect clinical decisions, thus improving patient care and safety.