For many children in the United States, school meals represent a vital source of reliable and nutritious food. Utilizing variation caused by the Community Eligibility Provision (CEP) in Georgia schools, we estimate models of school-level child health measured by the percentage of healthy weight children and average Body Mass Index (BMI) score. CEP eligibility is used as an instrument for CEP participation and the percentage of students enrolled in free and reduced-price school lunches, as well as in the reduced form. We find that CEP participation increases the percentage of healthy weight students in a school and reduces average BMI. We find no statistically significant evidence to support a deleterious effect from either the CEP or free school meals on child health outcomes. Subsample analyses suggest that that the effect of school meals on health varies across grade and location type, with no effect on high schools or rural schools.