There is little published data in relation to the effects of caffeine upon cycling performance, speed and power in trained cyclists, especially during cycling of approximately 60 s duration. To address this, eight trained cyclists performed a 1 km time-trial on an electronically braked cycle ergometer under three conditions: after ingestion of 5 mg x kg-1 caffeine, after ingestion of a placebo, or a control condition. The three time-trials were performed in a randomized order and performance time, mean speed, mean power and peak power were determined. Caffeine ingestion resulted in improved performance time (caffeine vs. placebo vs. control: 71.1 +/- 2.0 vs. 73.4 +/- 2.3 vs. 73.3 +/- 2.7 s; P = 0.02; mean +/- s). This change represented a 3.1% (95% confidence interval: 0.7-5.6) improvement compared with the placebo condition. Mean speed was also higher in the caffeine than placebo and control conditions (caffeine vs. placebo vs. control: 50.7 +/- 1.4 vs. 49.1 +/- 1.5 vs. 49.2 +/- 1.7 km x h-1; P = 0.0005). Mean power increased after caffeine ingestion (caffeine vs. placebo vs. control: 523 +/- 43 vs. 505 +/- 46 vs. 504 +/- 38 W; P = 0.007). Peak power also increased from 864 +/- 107 W (placebo) and 830 +/- 87 W (control) to 940 +/- 83 W after caffeine ingestion (P = 0.027). These results provide support for previous research that found improved performance after caffeine ingestion during short-duration high-intensity exercise. The magnitude of the improvements observed in our study could be due to our use of sport-specific ergometry, a tablet form and trained participants.