Special purpose finite element codes have been developed to such a userfriendly stage that application of these programs is no longer restricted to a small number of highly skilled specialists. Consequently these codes have been increasingly applied to obtain solutions to practical geotechnical problems. Although in general desirable, there are some risks associated with these developments because less experienced engineers who may not have the required theoretical background for performing complex nonlinear calculations are also tempted to use these programs. Additional problems arise from the fact that numerical analyses in geotechnical engineering have to address some very specific computational aspects (e.g. in situ stresses, excavation sequences, pore water pressures) which have no importance in other disciplines and which are sometimes difficult to handle by unexperienced users. Hence there is a strong need to develop benchmark tests with special emphasis on geotechnical engineering in order to provide some means for checking the reliability of the available software and formulate guidelines for using these programs.