We review recent advances towards the computation of string couplings. Duality symmetry, mirror symmetry, Picard-Fuchs equations, etc. are some of the tools.One of the main topics of this conference was the matrix model approach to noncritical strings. There the outstanding open problem is to go above c = 1. Here we want to review some recent progress in the 'old-fashioned' formulation of critical string theory with (c,c) = (15, 26) (in the case of the heterotic string). Since the description of the space-time degrees of freedom only uses up (6, 4) units of the central charge, one uses the remaining (9, 22) to describe internal degrees of freedom (gauge symmetries). We will not discuss any of the conditions which have to be imposed on the string vacua, such as absence of tachyons, modular invariance, etc. In the class of models we will mainly be concerned with, namely Calabi-Yau compactifications [1], they are all satisfied. We will rather address the problem of how to close the gap between the formal description and classification of string vacua and their possible role in a realistic description of particle physics. Even if one finds a model with the desired particle content and gauge symmetry, one is still confronted with the problem of computing the couplings, which determine masses, mixing angles, patterns of symmetry breaking etc. These couplings will depend on the moduli of the string model, which, in the conformal field theory language, correspond to the exactly marginal operators, or, in the Calabi-Yau context, to the harmonic (1,1) and * Partially supported by the Deutsche Forschungsgemeinschaft. † Contribution to the 'International Conference on Modern Problems in Quantum Field Theorie, Strings and Quantum Gravity', Kiev, June 8-17,1992.-1-