<p>Since its inception, Generative Artificial Intelligence (GenAI) has attracted significant attention in education and shows great potential in improving students’ academic performance. Although many studies suggest that academic performance is influenced by multiple factors, how these factors affect academic performance in GenAI-supported learning remains unclear. This study collected self-reported data from 40 students in a GenAI-supported learning context, including their AI literacy, perception of technology, student engagement, and academic performance. Using a configurational approach, methods such as fuzzy-set Qualitative Comparative Analysis (fsQCA) and Necessary Condition Analysis (NCA) were applied to explore the complex relationships among these variables and their synergistic effects on academic performance. The results show that no single factor is necessary for academic performance; instead, it arises from the interaction of multiple factors. Specifically, three different configurations were found to lead to high academic performance, all sharing high emotional engagement as a common feature. Moreover, complementary effects were observed among AI literacy, perception of technology, cognitive engagement, emotional engagement, and behavioral engagement, and their synergy significantly enhanced academic performance. The study also identified two configurations associated with low academic performance, highlighting the complexity and asymmetry of factors influencing learning outcomes. This research contributes to deepen the understanding of the multiple factors affecting academic performance in GenAI-supported learning, enriches social cognitive theory, and reveals the key role of emotional engagement in achieving high academic performance. Moreover, it offers several practical implications for educators on how to support students effectively in GenAI-supported learning.</p>