The antiproliferative properties of a group of 13 structurally diverse gold(III) compounds, including six mononuclear gold(III) complexes, five dinuclear oxo-bridged gold(III) complexes, and two organogold(III) compounds, toward several human tumor cell lines were evaluated in vitro using a systematic screening strategy. Initially all compounds were tested against a panel of 12 human tumor cell lines, and the best performers were tested against a larger 36-cell-line panel. Very pronounced antiproliferative properties were highlighted in most cases, with cytotoxic potencies commonly falling in the low micromolar-and even nanomolar-range. Overall, good-to-excellent tumor selectivity was established for at least seven compounds, making them particularly attractive for further pharmacological evaluation. Compare analysis suggested that the observed antiproliferative effects are caused by a variety of molecular mechanisms, in most cases "DNA-independent,'' and completely different from those of platinum drugs. Remarkably, some new biomolecular systems such as histone deacetylase, protein kinase C/staurosporine, mammalian target of rapamycin/rapamycin, and cyclin-dependent kinases were proposed for the first time as likely biochemical targets for the gold(III) species investigated. The results conclusively qualify gold(III) compounds as a promising class of cytotoxic agents, of outstanding interest for cancer treatment, while providing initial insight into their modes of action.