Gravitational lensing is a tool uniquely suited to studying the dark sector. Because its observable properties depend on all mass between the observer and source, regardless of how that mass interacts with light, lensing enables the study of total mass distributions from subplanetary to cosmological scales. To interpret strong lensing observations in the context of dark sector physics, they must typically be compared to cosmological simulations. This comparison is particularly challenging for the cluster strong lensing problem due to the simultaneous requirements for large simulation volume and high resolution, and the large modeling effort required for each lensing system. As a result, most existing constraints from strong lensing rely on analytic approximations, or theoretical predictions that do not account for effects such as line-of-sight structure, selection effects, or the systematics of mass modeling in observations. In this talk I will discuss how we can bridge the gap between observations and cosmological simulations, using a particular tension in the abundance of galaxy-galaxy strong lenses in cluster fields as an illustrative example. I will then discuss what the future holds for dark sector inference as JWST, Euclid, LSST, and Roman revolutionize both the resolution and statistical power of strong lensing observables.