Using Young Lives longitudinal data from Peru, this paper explores the relationship between socioeconomic status (SES) measured at the age of one, opportunities to learn (OTL) and achievement in mathematics ten years later. Four variables of OTL were measured: hours of class per year, curriculum coverage, quality of teachers' feedback, and level of cognitive demand. The last three were measured through an analysis of the exercises attempted by fourth grade students in their notebooks and workbooks. Multivariate analysis showed a robust association of one of the OTL variables (curriculum coverage, more specifically number of exercises attempted by students) with achievement in mathematics. Moreover SES at the age of one was significantly associated with this variable and with achievement by the time students were ten years old. Overall, the findings of the paper illustrate a highly unequal education system in which relatively poor children have fewer OTL in school.