Exact analytical benchmark examples in topology optimization may appear to be an academic exercise, but they are the only reliable basis for checking on the validity, convergence and accuracy of numerical methods. Michell’s (1904) truss topology optimization theory has not been extended from a single load design to multi-load stress-based exact optimization for over a century, although some multi-load compliance-based design problems were solved in the early nineties (Rozvany 1992, Rozvany, Zhou and Birker 1993) This significant gap in our knowledge is being filled by a current research project, which is going to be reviewed in this paper. First stress-based multi-load optimal elastic and plastic design will be compared, using also illustrative examples. ‘Plastic design’ refers to linearly elastic-ideal plastic structures, which can be designed by considering equilibrium conditions only, if we use the lower bound theorem of plasticity. Stress-based design considering different permissible stresses in tension and compression will also be discussed, and multi-load truss optimization with stability constraints considered. Displacement-based exact multi-load topology optimization will be the next topic looked at, and its results will be compared with stress-based examples. The first exact probabilistic structural topology optimization benchmark example (with a compliance constraint) was published by Rozvany and Maute (2011), and its extension to two stochastic load conditions was proposed by Suzuki and Haftka (2014). In this paper, some highly controversial recent allegations in the literature will be critically examined and refuted. It is to be remarked that at the ISSMO congress in Orlando only preliminary results of multi-load optimization were presented by the authors. The research results in this paper have not appeared yet in any other publication.