We first present a brief history of the origins of the complexity sciences as they have developed with respect to theories and models of cities. Our argument that the original view that cities might be understood and modelled as general systems from the top down, where the focus is on simulating the system in equilibrium, has radically shifted to viewing such systems as being continually out-of-equilibrium with a dynamic driven from the bottom up. This has important implications for how we model cities and we present the key features of a new generation of city models built around these ideas. We set this argument in the much wider context of changes in our perceptions about how we should plan cities. We note that the development of collaborative planning theories that see planning as dialogue between conflicting actors is in turn being influenced by conceptions of complexity. In particular, we argue that this shift from static to dynamic, top–down to bottom–up, is problematic for traditional notions of the optimum city which is inevitably an equilibrium to be aspired to. We conclude with some reflections about the difficulties for prediction that are intrinsic to this view of cities and their planning.