We report results of a large computational 'alloy by design' study, in which the 'chemical composition-mechanical strength' space is explored for austenitic, ferritic and martensitic creep resistant steels. The approach used allows simultaneously optimization of alloy composition and processing parameters based on the integration of thermodynamic, thermo-kinetics and a genetic algorithm optimization route. The nature of the optimisation depends on both the intended matrix (ferritic, martensitic or austenitic) and the desired precipitation family. The models are validated by analysing reported strengths of existing steels. All newly designed alloys are predicted to outperform existing high end reference grades.