In cycling, critical power (CP) and work above CP (W’) can be estimated through linear and nonlinear models. Despite the concept of CP representing the upper boundary of sustainable exercise, overestimations may be made as the models possess inherent limitations and the protocol design is not always appropriate. Objectives: To measure and compare CP and W’ through the exponential (CPexp), 3-parameter hyperbolic (CP3-hyp), 2-parameter hyperbolic (CP2-hyp), linear (CPlinear), and linear 1/time (CP1/time) models, using different combinations of TTE trials of different durations (approximately 1–20 min). Design: Repeated measures. Methods: Thirteen healthy young cyclists (26 ± 3 years; 69.0 ± 9.2 kg; 174 ± 10 cm; 60.4 ± 5.9 mL kg−1 min−1) performed five TTE trials on separate days. CP and W’ were modeled using two, three, four, and/or five trials. All models were compared against a criterion method (CP3-hyp with five trials; confirmed using the leaving-one-out cross-validation analysis) using smallest worthwhile change (SWC) and concordance correlation coefficient (CCC) analyses. Results: CP was considerably overestimated when only trials lasting less than 10 min were included, independent of the mathematical model used. Following CCC analysis, a number of alternative methods were able to predict our criterion method with almost a perfect agreement. However, the application of other common approaches resulted in an overestimation of CP and underestimation of W’, typically these methods only included TTE trials lasting less than 12 min. Conclusions: Estimations from CP3-hyp were found to be the most accurate, independently of TTE range. Models that include two trials between 12 and 20 min provide good agreement with the criterion method (for both CP and W’).