This paper proposes InqEduAgent, an agent model based on a large-scale language model (LLM) for effective learning partner matching in inquiry-based education. InqEduAgent analyzes learners' prior knowledge patterns and matches them with optimal learning partners through a generative agent that captures learners' cognitive and evaluative characteristics in real-world learning environments and an adaptive matching algorithm utilizing Gaussian process augmentation. Experimental results show that InqEduAgent performs optimally in various knowledge-learning scenarios and in the LLM environment. This contributes to the intelligent assignment of human-based learning partners and the design of AI-based learning partners. The code, data, and appendices are publicly available.
