Although reliability correlations of traditional power indices of the Wingate test have been well documented, no study has analyzed new generation power indices based on milliseconds obtained from a Peak Bike. The purpose of this study was to investigate the retest reliability of new generation power indices. Thirty-two well-trained male athletes who were specialized in basketball, football, tennis, or track and field volunteered to take part in the study (age: 24.3 +/- 2.2 years; body mass: 77 +/- 8.3 kg; height: 180.3 +/- 6.3 cm). Participants performed two Wingate all-out sessions on two separate days. Intra-class correlation coefficient (ICC), standard error measurement (SEM), smallest real differences (SRD) and coefficient of variation (CV) scores were analyzed based on the test and retest data. Reliability results of traditional power indices calculated based on 5-s means such as peak power, average power, power drop, and fatigue index ratio were similar with the previous findings in literature (ICC 0.94; CV 2.8%; SEM 12.28; SRD% 7.7%). New generation power indices such as peak power, average power, lowest power, power drop, fatigue index, power decline, maximum speed as rpm, and amount of total energy expenditure demonstrated high reliability (ICC 0.94; CV 4.3%; SEM 10.36; SRD% 8.8%). Time to peak power, time at maximum speed, and power at maximum speed showed a moderate level of reliability (ICC 0.73; CV 8.9%; SEM 63.01; SRD% 22.4%). The results of this study indicate that reliability correlations and SRD% of new generation power and fatigue-related indices are similar with traditional 5-s means. However, new time-related indices are very sensitive and moderately reliable.