Academic grit, known as passion and perseverance for long-term academic goals, has been widely studied in the educational context. However, few are the studies that analyze this variable as a specific domain and longitudinally. We used a sample of 4,853 students, assessing their levels of academic grit at two timepoints four years apart: in the fourth year (M = 9.9 years, SD = 0.41), and eighth year (M = 13.87 years, SD = 0.82) of compulsory education. We also evaluated their academic performance via their school grades in Spanish language and mathematics at three timepoints. Latent Class Analysis was used to identify underlying groups, and Latent Transition Analysis to examine the transitions between the latent groups at the different timepoints. We also performed a repeated measure ANOVA to analyse the influence of academic grit on school performance. We identified three groups, differentiated by the level of academic grit: Gritty, Industrious, and Careless group. Over time (between 10 and 14 years old) we saw a clear transition of students towards the groups with lower levels of grit. School performance decreased between the ages of 10 and 14, although not in a straight line, and the change was modulated by the level of academic grit. Promoting academic grit from the primary stage could dampen school failure in later stages.