To establish the test–retest reliability of pulmonary oxygen uptake (V̇O2), muscle deoxygenation (deoxy[haem]) and tissue oxygen saturation (StO2) kinetics in youth elite-cyclists. From baseline pedalling, 15 youth cyclists completed 6-min step transitions to a moderate- and heavy-intensity work rate separated by 8 min of baseline cycling. The protocol was repeated after 1 h of passive rest. V̇O2 was measured breath-by-breath alongside deoxy[haem] and StO2 of the vastus lateralis by near-infrared spectroscopy. Reliability was assessed using 95% limits of agreement (LoA), the typical error (TE) and the intraclass correlation coefficient (ICC). During moderate- and heavy-intensity step cycling, TEs for the amplitude, time delay and time constant ranged between 3.5–21.9% and 3.9–12.1% for V̇O2 and between 6.6–13.7% and 3.5–10.4% for deoxy[haem], respectively. The 95% confidence interval for estimating the kinetic parameters significantly improved for ensemble-averaged transitions of V̇O2 (p < 0.01) but not for deoxy[haem]. For StO2, the TEs for the baseline, end-exercise and the rate of deoxygenation were 1.0–42.5% and 1.1–5.5% during moderate- and heavy-intensity exercise, respectively. The ICC ranged from 0.81 to 0.99 for all measures. Test–retest reliability data provide limits within which changes in V̇O2, deoxy[haem] and StO2 kinetics may be interpreted with confidence in youth athletes.