Objectives To examine the association between education and mortality by age and gender in Chile. Methods We drew data from 10,147 adults aged 40 years and over from the Chilean Social Protection Survey linked to the national death records. We specify five Cox regression models to estimate the main and interaction effects for education levels, age, sex, and mortality. Results The hazard ratios reduce with increasing education. Secondary and tertiary education levels are associated, respectively, with 34% and 41% lower hazard rates than 0-4 years of schooling. Also, the educational gradient in mortality is significantly weaker at older ages, and it does not differ by gender. Discussion Our findings endorse the negative association of formal education with adult mortality, uncovering Chile's structural problems despite the economic improvements underway. It also corroborates the importance of novel longitudinal data for mortality analyses in Latin America.