This study aims to compare maximal lactate accumulation rate (V̇ Lamax) and power output (Pmax) between cycling and running in terms of reliability, differences between, and correlations among modalities. Eighteen competitive triathletes performed a 15-s all-out exercise test in cycling and a 100-m sprint test in running. Each test was performed twice and separated by one week. Exercise tests in cycling were performed on an ergometer whereas sprint tests in running were performed on an indoor track. Differences between trials and exercise modality were analyzed using two-way ANOVA. V̇ Lamax (ICC=0.894, ICC=0.868) and Pmax (ICC=0.907, ICC=0.965) attained ‘good’ to ‘excellent’ reliability in both cycling and running, respectively. V̇ Lamax was higher in running (d=0.709, p=0.016) whereas Pmax was lower in running (d=−0.862, p < 0.001). For V̇ Lamax, limits of agreement between modalities ranged from −0.224 to +0.437 mmol·l-1·s-1. Pmax correlated between modalities (r=0.811, p < 0.001), whereas no correlation was found in V̇ Lamax (r=0.418, p=0.084). V̇ Lamax is highly reliable in both modalities and higher in running compared to cycling. Since V̇ Lamax does not correlate between cycling and running, it should be determined sport-specifically.