Critical power (CP) conceptually represents the highest power output (PO) at physiological steady-state. In cycling exercise, CP is traditionally derived from the hyperbolic relationship of ∼5 time-to-exhaustion trials (TTE) (CPHYP). Recently, a 3-min all-out test (CP3MIN) has been proposed for estimation of CP as well the maximal lactate steady-state (MLSS). The aim of this study was to compare the POs derived from CPHYP, CP3MIN, and MLSS, and the oxygen uptake and blood lactate concentrations at MLSS. Thirteen healthy young subjects (age, 26 ± 3years; mass, 69.0 ± 9.2 kg; height, 174 ± 10 cm; maximal oxygen uptake, 60.4 ± 5.9 mL·kg−1·min−1) were tested. CPHYP was estimated from 5 TTE. CP3MIN was calculated as the mean PO during the last 30 s of a 3-min all-out test. MLSS was the highest PO during a 30-min ride where the variation in blood lactate concentration was ≤ 1.0 mmol·L−1 during the last 20 min. PO at MLSS (233 ± 41 W; coefficient of variation (CoV), 18%) was lower than CPHYP (253 ± 44 W; CoV, 17%) and CP3MIN (250 ± 51 W; CoV, 20%) (p < 0.05). Limits of agreement (LOA) from Bland–Altman plots between CPHYP and CP3MIN (–39 to 31 W), and CP3MIN and MLSS (–29 to 62 W) were wide, whereas CPHYP and MLSS presented the narrowest LOA (–7 to 48 W). MLSS yielded not only the maximum PO of stable blood lactate concentration, but also stable oxygen uptake. In conclusion, POs associated to CPHYP and CP3MIN were larger than those observed during MLSS rides. Although CPHYP and CP3MIN were not different, the wide LOA between these 2 tests and the discrepancy with PO at MLSS questions the ability of CP measures to determine the maximal physiological steady-state.