The most recent experiments of disruption mitigation by massive gas injection in ASDEX Upgrade have concentrated on small-relatively to the past-quantities of noble gas injected, and on the search for the minimum amount of gas necessary for the mitigation of the thermal loads on the divertor and for a significant reduction of the vertical force during the current quench. A scenario for the generation of a long-lived runaway electron beam has been established; this allows the study of runaway current dissipation by moderate quantities of argon injected. This paper presents these recent results and discusses them in the more general context of physical models and extrapolation, and of the open questions, relevant for the realization of the ITER disruption mitigation system.