Precise time extrapolation of creep rupture data is a requirement for components operating at high temperatures. Although qualified models have been available for many years, there has been no way of quantifying the magnitude of the error in the results. To make an error analysis meaningful, requirements on the first and second derivatives are introduced, which are taken into account in the optimization of the model of the predicted creep rupture values. Formulae for the error are presented both for the case when the extrapolated values are interpolated from a master curve and when the values are extrapolated from the master curve. The criteria have been applied with a timeerature parameter for the austenitic creep resistant stainless steel Super304H (18Cr9Ni3CuNbN). The results showed a stable behavior with limited oscillating of the creep rupture curves. The prediction automatically fulfilled the post-assessments tests (PATs) proposed by the European Collaborative Creep Committee (ECCC), which are considered to be quite severe requirements.