Beijing – At the 2025 World Agrifood Innovation Conference (WAFI 2025) in Beijing, China Agricultural University (CAU) introduced Shennong Large Model 3.0, a next-generation AI system designed to make smart farming more accessible and efficient worldwide.
According to lead researcher Wang Yaojun, the new model slashes computational demands while boosting performance by 5% over its predecessor. The architecture was carefully redesigned to strike an optimal balance between a lightweight footprint and high-throughput processing.
Building on the Shennong lineage, version 1.0 laid the groundwork with core agricultural knowledge and question-answering, while 2.0 added multimodal functions and expanded sector use. Now, Shennong 3.0 takes another leap forward.
Alongside the AI core, CAU rolled out an agent platform featuring 36 scenario-specific agents across six categories—from smart breeding to precision farming. These agents integrate seamlessly with sensors and machinery, delivering tailored guidance on planting cycles, pest control, and crop management.
Pilot programs are already live in Beijing, the Inner Mongolia Autonomous Region, and Heilongjiang Province, offering localized plant protection and customized cultivation advice for farmers in diverse settings.
Trained on a specialized agricultural dataset—including over 10 million knowledge graphs, 50 million production records, and 20,000 monographs—Shennong 3.0 is primed for scalable deployments in both developed and emerging markets.
As global food security and sustainability pressures rise, Shennong 3.0 represents a leap toward intelligent, data-driven farming that's easy to deploy and customize, laying new groundwork for the future of agriculture.
Reference(s):
cgtn.com
