Galaxy clusters result from the gravitational collapse of rare high peaks in the initial density field. Their characteristic formation time is comparable to the Hubble time, which makes them sensitive to the overall growth of structure in the universe. The abundance of galaxy clusters thus provides valuable information on the background cosmology, law of gravity and initial conditions. Assuming Gaussian initial conditions and in the limit of ΛCDM background evolution, f(R) gravity can still alter the formation of structures compared to standard gravity in interesting and measurable ways. I will present current conservative constraints on a viable class of f(R) theories from the observed abundance of X-ray selected massive galaxy clusters out to z ~ 0.5. I will also show that an improved mass function model calibrated on state- of-the-art cosmological simulations can realise the full potential of ongoing and upcoming cluster surveys, restricting departures from General Relativity on cosmic scales to the level allowed by local tests of gravity.