We investigated the measurement of insulin sensitivity (S I) with a standardized hyperglucidic breakfast (SHB) compared to minimal model analysis of an intravenous glucose tolerance test (S I-IVGTT) in 17 patients clinically referred as type 2 diabetics, not yet treated by insulin, and representing a wide range of body mass index and S I. To classify the patients, ten meal-tolerance test-based calculations of S I (MTT-S I) were compared to S I-IVGTT, and their reference values and distribution were measured on a separate sample of 200 control SHBs and 209 control IVGTTs. Eight MTT-SI indices exhibit significant correlations with S I-IVGTT: Mari's OGIS index, BIGTT-SI|0-30-120, BIGTT-SI|0-60-120, 1/G b I m, Caumo's oral minimal model (OMM), Sluiter's index "A" = 10(4)/(I p·G p), Matsuda's composite index given by the formula ISIcomp = 10(4)/(I b G b I m G m)(0.5), S I = 1/I b G b I m G m with r (2) ranging between 0,53 and 0,28. S I-IVGTT and S I-MTT exhibited in the lower range a very different (non-normal) pattern of distribution and thus the cutoff value for defining insulin resistance varied among indices. With such cutoffs, S I-MTT < 6.3 min(-1)/(μU/ml) 10(-4) with Caumo's OMM was the best predictor of insulin resistance defined as S I-IVGTT < 2 min(-1)/(μU/ml) 10(-4). Other indices, including OGIS and BIGTT, resulted in more misclassifications of patients. HOMA-IR and QUICKI were poor predictors. The formula [Formula: see text] satisfactorily predicts IVGTT-derived glucose effectiveness in type 2 diabetics. Thus, SHB appears suitable for the measurement of S I and S G in type 2 diabetics, and the OMM seems to provide the most accurate SHB-derived index in this population.