Continuous-time q-learning introduces integrated q-functions for mean-field control problems, enabling efficient learning algorithms.