On January 13, 2026, Nanjing Agricultural University made a significant announcement by officially launching 'Sinong'—China's inaugural open-source, vertical large language model tailored specifically for the broad agricultural industry. Drawing on the university's strong disciplinary expertise, the model has aggregated an impressive collection of over 4 billion pieces of agricultural professional data. This extensive dataset forms the backbone of the model, encompassing close to 9,000 books, more than 200,000 research papers, and a wide range of policy standards. To bolster its understanding and reasoning abilities in tackling intricate agricultural challenges, the model integrates multi-dimensional data, including chain-of-thought reasoning and contextual references. Moreover, it employs a multi-agent retrieval-augmented framework to guarantee the precision and currency of the content it generates. At present, 'Sinong' is readily accessible and fully open-source on platforms like ModelScope and GitHub, offering versions with 8 billion (8B) and 32 billion (32B) parameters.