Accurately modeling neutrino flavor oscillations in global simulations of core-collapse supernovae or neutron star mergers remains a major challenge, albeit a potentially crucial one for making reliable predictions. Indeed, it is now widely recognized that flavor instabilities, in which neutrino distributions computed in classical transport frameworks can be dramatically reshaped by flavor conversion, are ubiquitous in such environments.
In this talk, I will describe recent progress in understanding and modeling these instabilities, focusing on the development of subgrid models of neutrino flavor transformation. In this approach, insights from detailed studies of local quantum neutrino transport are used to formulate effective prescriptions that modify classical neutrino distributions, enabling their incorporation into large-scale simulations. Over the past year, the first supernova and neutron star merger simulations including these subgrid treatments have been performed, a key step toward fully incorporating neutrino physics into such models and assessing its wide-ranging implications.