The Relative Biological Effectiveness (RBE) of 8 fast-neutron beams, 5 proton beams and 1 carbonion beam was determined using as biological criterion intestinal crypt regeneration in mice, i.e. an in vivo system. These beams are used or planned for clinical cancer therapy applications. In addition, the RBE of 6 epithermal neutron beams, used or planned for Boron Neutron Capture Therapy (BNCT), was determined; no boron was administered. The goal of the program was to improve the exchange of information between the centers, facilitate the interpretation of the results and increase the safety of the clinical applications. In all visited centers, the same technique was applied in the same conditions by the same radiobiology team. The number of regenerating crypts per circumference was scored 3.5 days after single fraction total body irradiation. The control irradiations were performed locally using cobalt-60 units. The mice were randomized according to radiation quality and dose level. (1) For fast neutron beams, the RBE (Ref. cobalt-60 gamma rays) increases with decreasing energy (from approximately 1.7 for p(65)+Be neutrons to approximately 2.4 for d(14.5)+Be neutrons). In addition, it is specific to each facility and depends on the nuclear reaction (p or d + Be), target and collimation type. (2) For proton beams, the RBEs (Ref cobalt-60 gamma rays) at the reference position (middle of a 7-cm Spread Out Bragg Peak, SOBP) range between 1.08 and 1.18. They might differ by approximately 6-8% according to the mode of beam production or delivery. The RBEs at the end of the SOBP are always 5-10 % higher than at the middle of the SOBP. (3) For the carbon ion beam studied at NIRS in Chiba, Japan, the RBE significantly increases with depth. Relative to gamma rays, it ranges from 1.3 in the initial plateau, 1.6 at the beginning, 1.7 at the middle and 1.9 at the end of a 6-cm SOBP. 4) In BNCT beams, the radiation quality (in particular the relative contribution of the different dose components) varies rapidly with depth and depends strongly on the arrangement of the irradiation set-up (e.g. presence or not of back scattering material). Moreover, the (total) dose rates are highly variable (from 0.05 to approximately 0.5 Gy/min) according to the power of the reactors. Wide range of RBE values (Ref. gamma rays) was thus obtained (RBE = 1.4 - 2.2) at shallow depths of 1.5 - 2.5 cm. Intestinal crypt regeneration in mice is an in vivo system perfectly suitable to perform intercomparisons between centers applying different types of non-conventional radiation qualities. It was proven to be reproducible, reliable and accurate, and becomes progressively recognized worldwide as part of the Quality Control (QA) procedures for new beams. It should be stressed that the observed RBE for intestinal crypt cells after a single high dose provide some radiobiological characterization of the radiation quality but cannot be used as the RBE weighting factor in clinical prescriptions.