Flexible problem solving consists of the dynamic selection and configuration of problem solving methods for a particular problem type, depending on the particular problem and the goal of problem solving. In this paper, we propose an architecture that supports such flexible problem solving automatically. For this purpose, problem solving methods are described in a uniform way, by an abstract model of components, which together define the functionality of the methods. Such an abstract model is used for dynamic selection and configuration of the problem solving methods. The proposed architecture for flexible problem solving consists of well known reflection techniques: two object-meta relations, a definable naming mechanism and the axiomhood and theoremhood reflection rules. We have succeeded in using standard meta-architecture techniques to enable flexible problem solving.