Abstract
This study comprehensively analyses ten advanced generative AI models including GPT-4, Microsoft Copilot, Claude, DeepSeek, Pi and others to assess their applicability in higher education for computer science students. The study uses a mixed-method approach involving qualitative and quantitative analyses to evaluate these models across four key groups: architecture, content quality, adaptability, and performance. The results show that while each model has certain strengths - such as GPT-4’s content creation capabilities and Pi’s adaptability - none is the optimal choice across all clusters. The study emphasizes the importance of aligning the choice of AI tool with specific educational goals and needs. It also emphasizes the need for continuous evaluation of AI technologies to ensure their effectiveness in dynamic educational environments. The study contributes to the growing discourse on AI in education by offering a sound framework for evaluating AI models and guiding their implementation in educational environments.