The purpose of this study is to compare and examine the results of speech analysis in the classroom- specific model conducted by Yasumori (2024b) with the results of speech analysis using a new general- purpose large language model (LLM). The research method involves conducting two analyses—one using a lesson-specific model and the other using a general-purpose large language model (LLM)—on the utterances made by an in-service teacher during a single mathematics lesson, and comparing the results. The results of the study indicate that analysis using a general-purpose large language model (LLM) demonstrates higher validity than the lesson-specific model proposed by Yasumori (2024b).
