Evidence from cognitive neuroscience suggests that learning counterintuitive concepts in mathematics and science requires inhibitory control (IC). This prevents interference from misleading perceptual cues and naïve theories children have built from their experiences of the world. Here, we (1) investigate associations between IC, counterintuitive reasoning, and academic achievement and (2) evaluate a classroom-based computerised intervention, called Stop & Think, designed to embed IC training within the learning domain (i.e. mathematics and science content from the school curricula). Cross-sectional analyses of data from 627 children in Years 3 and 5 (7-to 10-year-olds) demonstrated that IC, measured on a Stroop-like task, was associated with counterintuitive reasoning and mathematics and science achievement. A subsample (n = 456) participated either in Stop & Think as a whole-class activity (teacher-led, STT) or using individual computers (pupil-led, STP), or had teaching as usual (TAU). For Year 3 children (but not Year 5), Stop & Think led to better counterintuitive reasoning (i.e. near transfer) in STT (p < .001, η p 2 = .067) and STP (p < .01, η p 2 = .041) compared to TAU. Achievement data was not available for Year 3 STP or Year 5 STT. For Year 3, STT led to better science achievement (i.e. far transfer) compared to TAU (p < .05, η p 2 = .077). There was no transfer to the Stroop-like measure of IC. Overall, these findings support the idea that IC may contribute to counterintuitive reasoning and mathematics and science achievement. Further, we provide preliminary evidence of a domain-specific IC intervention with transferable benefits to academic achievement for Year 3 children.