The Wendt model activated via G4NEUTRON_HP_USE_WENDT_FISSION_MODEL=1 seems not to be thread-safe. If run in single thread it works, but in MT mode the data members
RemainingA_
RemainingZ_
are obviously accessed by several threads and they assume negative values, which leads to a crash. Also, it should be checked somewhere that this model and G4NEUTRONHP_PRODUCE_FISSION_FRAGMENTS=1 are not active at the same time since they are mutually exclusive.