Participants completed a before/after and a similar/different relational task, using the Implicit Relational Assessment Procedure (IRA P), and subsequently took the Kaufman Brief Intelligence Test (K-BIT). For each relational task, response latencies were measured first on consistent trials, where participants responded in accordance with preestablished verbal relations, and then on inconsistent trials, where they responded against these relations. A difference-score was calculated by subtracting consistent from inconsistent response latencies. The inconsistent trials and the difference-score provided measures of relational flexibility. Results showed that faster responding on the IRAP and smaller difference-scores predicted higher IQ. These findings suggest that relational flexibility is an important component of intelligence and might therefore be targeted in educational settings.