<p>Studies have investigated the effect of cognitive activation on students’ performance, with some suggesting positive effects, while others report mixed or negative ones. This study aims to explore the factors contributing to this inconsistency by investigating the association between students’ perceived cognitive activation and their mathematics performance when considering students’ grades, performance levels, and cognitive activation intensity. A large sample of 28,726 students from the Local Basic Education Quality Assessment in China dataset was analyzed in this study, including 17,378 fourth graders and 11,348 eighth graders. The results revealed that, overall, there was a stronger association between basic cognitive activation and mathematics than the association between deep cognitive activation and mathematics for both fourth- and eighth-grade students. When considering grades and performance levels, eighth-grade students showed a stronger association between deep cognitive activation and mathematics compared to fourth-grade students. The association between basic cognitive activation and mathematics was stronger for students with low performance, while the association between deep cognitive activation and mathematics was stronger for students with high performance. This emphasizes the necessity of considering the grades and students’ performance levels when implementing cognitive activation. The implications for educators, particularly regarding the appropriate utilization of cognitive activation during educational practices, are discussed.</p>