Assessment of plasma insulin and glucose outcomes is important in paediatric studies aimed at reducing future risk of type 2 diabetes and cardiovascular disease. The aims of this study are to determine the between-method agreement and the day-to-day reliability of fasting and oral glucose tolerance test (OGTT)-derived estimates of insulin sensitivity and β-cell function in healthy boys. Fasting and OGTT assesments of insulin resistance and β-cell function were performed on 28 boys (12.3±2.9 years). Measurements were repeated after 1 week (fasting, n=28) and 1 day (OGTT, n=8). Agreement between estimates of insulin resistance and β-cell function was examined using Pearson's correlation coefficient. Reliability was assessed using change in the mean, Pearson's correlation coefficient, and typical error expressed as a coefficient of variation (CV). The Matsuda index was positively related with QUICKI (r=0.88, P<0.001) and negatively related to HOMA-IR (r=-0.76, P<0.001). The Cederholm index was not significantly related with fasting estimates of insulin resistance (all r<0.40, P>0.05). For reliability, QUICKI had the lowest CV% for the fasting (4.7%) and the Cederholm index for the OGTT (6.4%) estimates. The largest CV% was observed in fasting insulin (30.8%) and insulinogenic index 30' (62.5%). This study highlights differences in between-method agreement and day-to-day reliability for estimates of insulin resistance in youth. The low CV supports the use of the FGIR (fasting) and Cederholm (OGTT) indices in this population.