This diagnostic/prognostic study evaluates the accuracy of a large language model against physician diagnoses in pediatric cases.