We conducted two empirical studies to (1) explore the latent structure of cognitive flexibility in children as measured by performance-based tasks, (2) analyze the contribution of working memory (WM) and inhibition to reactive and spontaneous flexibility, and (3) examine the contribution of the different flexibility components to academic skills (i.e., reading comprehension and writing) and creativity. In S1 ( = 112 8- to 12-year-old children), confirmatory factor analysis (CFA) showed best fit for the two-factor solution with reactive flexibility (RF) and spontaneous flexibility (SF) as separate but related components. When considering the joint contribution of WM and inhibition to both cognitive flexibility components, it was found that WM and inhibition contributed to SF, whereas only inhibition contributed to RF. Besides, only SF proved to be a significant predictor of writing and reading comprehension by using a latent-variable structural equation approach (SEM). In S2 ( = 177 8- to 13-year-old children), hierarchical regressions and SEM models showed consistently that the flexibility component in relation to creativity deals with the ability to generate diverse responses driven by internal stimuli (i.e., spontaneous flexibility). Taken together, these results suggest that cognitive flexibility is not a unified construct. Besides, the close relationship between SF and creativity and academic skills raise the question whether considering SF as a higher-level form of cognitive flexibility (different from a lower-level shifting skill) could be an interesting approach for the study of cognitive flexibility in children.