Analysis of International Competitive Landscape for Large Language Model Technologies Based on Patent Metrics

This paper explores the competitive landscape in the field of large language model technologies through the analysis of patent quantity. Patent data related to large language models from major global patent databases between 2004 and 2023 were collected to construct a patent value evaluation system based on three dimensions: technology, legal, and market. By comparing the patent application volume and patent quality of different companies and countries, the research intensity of their development activities in the field of large language models was evaluated.Preliminary results indicate that the competitive landscape in the field of large language model technologies is undergoing rapid changes. While some traditional technology giants possess a large number of patents in this field, there are also emerging start-ups and research institutions that are quickly gaining prominence. Additionally, this study has observed a rapid growth trend in the number of patent applications for large language model technologies in recent years, indicating an acceleration in research and development activities in this field.