This thesis explores the capabilities of freely available AI models, which can also be used offline, in generating code for website development. The research focuses on evaluating the performance of the five best models through the task of creating a simple responsive website based on a predefined design. Each model is assessed using six key criteria: HTML quality, CSS quality and visual fidelity, code efficiency and optimization, functionality and interactivity, semantic HTML and accessibility, and responsiveness. The theoretical part of this thesis examines the evolution of AI technologies, starting from early chatbot systems and AIML-based solutions that laid the foundation for modern AI applications to the rise of natural language processing and artificial intelligence. The practical part involves designing a simple responsive website and creating a task that will be assigned to each model accordingly. The research employs a standardized evaluation framework to ensure consistency and fairness across all models. The results highlight significant differences in model capabilities, emphasizing strengths and weaknesses. The goal of this research is to identify the best model for generating code for the development of a simple responsive website.