Gamification, which involves incorporating game design features into educational contexts, is frequently advocated as a means to rekindle student interest, the studies show evidence is varied and often overlooks the intricate dynamics present in actual classroom environments. This research investigated whether a 12-week gamified instructional approach utilizing elements such as points, badges, levels, collaborative quests, and a narrative framework enhanced academic performance, multidimensional engagement, and intrinsic motivation compared to traditional teaching methods of the same curriculum. Eight classes were evaluated, including two dedicated science classes (N = 312 enrolled; 304 analyzed), which were alternately assigned to either a gamified group (k = 4, n = 150) or a control group (k = 4, n = 154). The results were assessed using covariate-adjusted models with cluster-robust inference alongside class-level permutation tests and mediation analyses. After adjusting for pre-test scores, the gamified approach yielded better post-test results than the control condition (B = 4.58 points, 95% CI; Hedges' g = .59) and showed greater intrinsic motivation (g = .50). The difference in engagement levels was less pronounced (g = .26) and while statistically significant under conventional analysis, it did not hold up under cluster-robust scrutiny (p = .373). Engagement served as a partial mediator of the effect on academic achievement (indirect effect = .56, 95% bootstrap CI [.12, 1.15]), contributing approximately 12% to the overall impact. Although the advantages tended to be greater for students with lower prior achievements, this interaction lacked statistical significance. The findings suggest a prudent focus on mechanisms: gamification can enhance motivation and learning when designed with principles of autonomy, competence, and relatedness in mind; however, claims regarding engagement should be substantiated by larger class samples than those typically found in current research literature.