The present paper is predicated on the utilization of disparate Machine Learning Models in the standardized assessment tests “Aprender,” which were developed in Argentina, for the purpose of measuring Performance in Language and Mathematics. The models also collect information on conditioning factors that allow for the analysis of the contexts in which these learning processes take place. The present study proposes to implement this approach with data from the 2016 edition of the sixth-grade assessment in Elementary School. In the research stage, the Performance in Language and Mathematics was analyzed, and the results are presented in this document. To this end, a preliminary selection of features was conducted, followed by the selection of five Machine Learning Models. The model that attained the highest level of accuracy was identified as the optimal choice. Furthermore, the datasets utilized were subjected to a preliminary treatment, during which missing and negative data were completed using the mean of each respective column. The optimal results were determined based on the most effective calculation parameters, as determined by the evaluation tests. In each case, the most significant features that led to the best results were identified within the previously selected set of features. Finally, an analysis was conducted to assess the significance of these features across two performance groupings: “Satisfactory/Advanced” and “Basic/Below Basic”.