Chinese scientists have released the soybean vertical large model "Fengzhu" 2.0.

According to reports from Anhui Agricultural University, on August 16, at the 32nd National Soybean Research and Production Symposium held in Nanning, Guangxi, Professor Wang Xiaobo, representing the research and development team, announced the release of Fengzhu 2.0, a vertical large model in the field of soybeans. This model achieves multimodal data integration, reducing the bias of single model responses, and helps to promote data-assisted breeding for soybeans. Fengzhu 2.0 integrates soybean germplasm resources, genomic and coding sequences, proteins, transcripts, gene expression, disease samples, breeding experiment records, and both Chinese and foreign literature, forming a structured knowledge system for soybean research. Compared to the initial version, Fengzhu 2.0 introduces a new collaborative analysis mechanism among multiple models. According to the research and development team, Fengzhu 2.0 embeds certain interfaces for general large models, which can each answer the same professional question related to soybeans. The summary module of Fengzhu 2.0 then synthesizes these responses, integrating domain knowledge, proprietary data, and structured comparison rules to comprehensively analyze the consensus, differences, and evidence integrity within the responses of each model, resulting in clearer and more reliable professional conclusions that contribute to the stability of soybean expertise services.