Northwest A&F University, in collaboration with Huawei, has unveiled the first agricultural large model powered by domestically developed computing resources—the Jiurang Yunxing Agricultural Large Model.

From August 9 to 10, the Fourth Annual Academic Conference of the Smart Agriculture Professional Committee of the Chinese Society of Crop Science was held in the Yangling Demonstration Zone. During the conference, Northwest A&F University and Huawei Technologies Co., Ltd. jointly unveiled the Jiurang Yunxing Agricultural Large Model Joint Solution (hereinafter referred to as “Jiurang Yunxing”). This is the first domestically developed agricultural large model trained and released on a homegrown computing platform, marking an important step forward for China’s smart agriculture in achieving both independent controllability and advanced intelligence.

Northwest A&F University, in collaboration with Huawei, has unveiled the Jiurang Yunxing Agricultural Large Model 1.0.

Relying on the computing power of Ascend AI cloud services and the efficient model‑development capabilities of the ModelArts platform, Jiurang Yunxing has leveraged core technologies such as multi‑source data fusion, knowledge‑graph construction, and intelligent analytics and decision‑making to overcome the longstanding bottlenecks in traditional agricultural data processing—namely, data fragmentation, difficulties in integration, and low analytical efficiency.

Zhang Hongming, Dean of the School of Information Engineering at Northwest A&F University, stated that Jiurang Yunxing has deeply integrated a vast array of academic resources—covering master’s and doctoral dissertations, specialized textbooks, and frontline production data—and continues to innovate in its algorithmic architecture, enabling intelligent applications across disciplines and fields. Its precise matching and problem-solving capabilities allow farmers, when confronted with specific challenges such as apple pollination or genetic breeding, to swiftly obtain highly feasible solutions vetted by experts, rather than merely generic information found online.

Jiurang Yunxing has launched six specialized large-scale models across six key domains—fruit cultivation, animal husbandry, field crops, traditional Chinese medicinal herbs and tea, intelligent control, and soil conservation—providing scenario‑specific solutions to address industry‑wide pain points.

• ZhenGuo Large Model: Deep analysis of genetic and environmental data, enabling end-to-end intelligent solutions for breeding and cultivation, water and fertilizer management, as well as pest and disease monitoring and maturity assessment.

• HuiMu Large Model: Designed for livestock farming, it covers precision breeding, intelligent environmental control, disease early warning, and nutrition‑health monitoring, enabling quality improvement and efficiency gains across the entire production process.

• Fenghe Large Model: Focused on field crops, it integrates intelligent breeding, germplasm innovation, and real-time crop‑growth monitoring to drive simultaneous improvements in yield and quality.

• Qin Ultra-Large Model: Targeting specialty industries such as traditional Chinese medicinal materials and tea, it enables integrated solutions for intelligent production‑area matching, disaster prevention and control, precise determination of optimal harvest timing, and optimized harvesting‑price setting.

• Intelligent Control Large Model: Integrates real-time environmental analysis with dynamic equipment control, enabling precision management of water and fertilizer applications and precise pesticide application via drones and robots, thereby advancing the automation of agricultural production.

• Baorang Large-Scale Model: Designed for soil degradation management, it offers intelligent monitoring, analysis, and remediation solutions to safeguard farmland quality and production stability.

At present, Jiurang Yunxing has been piloted in Shaanxi, Shanxi, Sichuan, Gansu, Ningxia, Xinjiang, Inner Mongolia, and other regions, serving a cumulative cultivated area of over 200,000 mu and generating direct economic benefits exceeding RMB 100 million.

In collaboration with large-scale agribusinesses, the team has decomposed field‑level production processes into more than a hundred standardized steps, meticulously calculating inputs down to a few cents per mu, thereby achieving industrial‑grade precision in operations and control. This capability—rooted in real‑world scenarios and validated through large‑scale production—enables the large model to deliver exceptional stability and replicability in practical applications.

Zhang Hongming explained that the ultimate goal of agricultural intelligence is to achieve natural human–robot interaction, enabling agricultural robots to truly “understand human language.” Jiurang Yunxing serves as a crucial foundation for this vision: it not only comprehends natural-language commands from humans but also translates them into precise production execution plans, establishing a seamless intelligent workflow that bridges understanding and action.

Looking ahead, Northwest A&F University and Huawei will continue to deepen the integration of the Jiurang Yunxing agricultural large model with China’s core industrial systems, ensuring that technological innovation is fully implemented across the entire agricultural value chain and supporting rural revitalization and the modernization of agriculture. As Zhang Hongming emphasized, “The convergence of domestically developed computing power and agricultural large models represents not only a technological upgrade but also a pivotal step toward a comprehensive digital and intelligent transformation of the agricultural sector.”

Source: WeChat Official Account: AgriTech Hero – Digital and Smart Agriculture