It has been shown that the critical power (CP) in cycling estimated using a novel 3-min all-out protocol is reliable and closely matches the CP derived from conventional procedures. The purpose of this study was to assess the predictive validity of the all-out test CP estimate. We hypothesised that the all-out test CP would be significantly correlated with 16.1-km road time-trial (TT) performance and more strongly correlated with performance than the gas exchange threshold (GET), respiratory compensation point (RCP) and VO2 max. Ten club-level male cyclists (mean±SD: age 33.8±8.2 y, body mass 73.8±4.3 kg, VO2 max 60±4 ml·kg(-1)·min(-1)) performed a 10-mile road TT, a ramp incremental test to exhaustion, and two 3-min all-out tests, the first of which served as familiarisation. The 16.1-km TT performance (27.1±1.2 min) was significantly correlated with the CP (309±34 W; r = -0.83, P<0.01) and total work done during the all-out test (70.9±6.5 kJ; r = -0.86, P<0.01), the ramp incremental test peak power (433±30 W; r = -0.75, P<0.05) and the RCP (315±29 W; r = -0.68, P<0.05), but not with GET (151±32 W; r = -0.21) or the VO2 max (4.41±0.25 L·min(-1); r = -0.60). These data provide evidence for the predictive validity and practical performance relevance of the 3-min all-out test. The 3-min all-out test CP may represent a useful addition to the battery of tests employed by applied sport physiologists or coaches to track fitness and predict performance in atheletes.