This paper studies peer effects on individual performance among elite athletes in high-stakes tournaments. I ask whether the presence of a teammate affects athletes' performance using data from the latest seven World Swimming Championships. To identify causal effects I apply a regression discontinuity design by comparing athletes' performance in the finals when their teammate barely qualified and barely not qualified for the same final. Female athletes accompanied by a teammate swam 0.41%-0.56% of the average time faster, or ranked by 0.75-1.16 higher in the final. Male athletes' performance is unaffected. Potential channels and gender differences are discussed.