Abstract
This article is devoted to the use of generative AI to build 3D models of productive equipment and to implement this in digital twins. The main goal of the paper is to compare the opportunities of Meshy.ai, Masterpiece X, and Sloyd.ai for the automation of 3D modeling to provide an efficient way of generating complex objects. For the experiment, the MMT-ST-1000 chocolate tank model was used. The basis for comparison in the presented research work includes accuracy, detail, geometric integrity, and the number of polygons. Meshy.ai was found to be the most accurate in the case of images, but there were topological errors with a high degree of polygonality. Masterpiece X did not have sufficient detail due to the limited query text. Sloyd.ai works efficiently in the case of simple models because it must handle only a few polygons, though it might struggle hard to construct a complex object. Another fresh approach for improving accuracy can be to decompose a model into parts and then construct and refine the model through some 3D editor application to get detailed models with very good geometry without excessive computation. It narrows down to accelerating the creation of digital twin models and extending their use during production.