The minimal supersymmetric standard model is a popular and well-motivated extension of the standard model. As such, it has been constrained by a large number of different experimental searches. To truly assess the impacts of these experiments on the model one must perform a global fit, scanning over the multi-dimensional parameter space and combining all the data in a statistically rigorous manner. In this talk, I presented results from global fits of supersymmetric models performed with GAMBIT, the Global and Modular Beyond-the-Standard-Model (BSM) Inference Tool. I showed MSSM results from the latest GAMBIT papers, as well as exciting preliminary results from a dedicated study of the collider constraints on the electroweakino sector.