Аннотация:Nowadays, regulatory documents and models for assessing the quality of software are built on the basis of a paradigm that determines that the quality of programs will be checked for compliance with the terms of reference for their development. But, as practice has shown, such a paradigm does not fully correspond to modern conditions, providing not quality control, but verification of the compliance of programs with the customer’s expectations formulated at the initial stage of development. Based on the analysis of existing approaches to assessing the quality of software development, proposals have been synthesized to refine the paradigm of its assessment. The paper presents the formulation of a scientific and practical problem and proposes one of the approaches to its solution, based on the refinement of the currently used quality assessment paradigm, the transition from a rigid preset model to a model refined in the course of work. The solution to the formulated problem will provide a general increase in the efficiency of automated control due to clarification of the paradigm for quality assessment, transition to the use of a dynamic model for assessing the quality of software being developed