On August 30, at the 2026 Medical AI Innovation Forum in Shanghai, Ruijin Hospital and Huawei Cloud officially released the RuiPath 2.0 pathology AI model. The new model features comprehensive upgrades across diagnostic capabilities, interpretability, rare disease data training, performance, and lightweight deployment. It has already been deployed in over 90 hospitals across China, marking a significant step in the large-scale clinical adoption of pathology AI.
Five Core Upgrades: From "Functional" to "Practical"
According to Zhou Yuefeng, Director of Huawei Cloud, RuiPath 2.0 achieves five key breakthroughs based on real-world clinical needs:
1. Most Comprehensive Diagnostic Capabilities in the Industry
With 7 billion parameters, RuiPath 2.0 covers 19 common cancer types. Its diagnostic tasks have been expanded to 205, the most comprehensive in the industry, covering over 90% of pathologists' daily diagnostic scenarios with a practical accuracy of 90%.
2. First to Achieve Precise Interpretable Annotations
For the first time, the model provides precise and interpretable annotations for key pathological features including histological subtype, invasion depth, and mitotic count. This makes the AI's diagnostic process transparent and traceable, significantly boosting clinicians' trust.
3. Using AI to Solve the Rare Disease Data Bottleneck
To address the industry-wide challenge of scarce rare disease samples, Huawei Cloud and Ruijin Hospital pioneered the use of a pathology multimodal generative model to synthesize training data. This approach achieved 96.48% accuracy in diagnosing lymphoma.
4. Performance Surpassing International Benchmarks
Across 59 downstream diagnostic tasks, RuiPath 2.0 achieved state-of-the-art (SOTA) results in 42 tasks, outperforming leading international pathology foundation models such as Harvard's UNI2 and Microsoft's Virchow2.
5. Lightweight Deployment for Grassroots Accessibility
The lightweight version, RuiPath 2.0 Edge, has only 2 million parameters and can be directly deployed on standard hospital PCs without expensive GPU servers. It enables minute-level inference in collaboration with the cloud, assisting with 90% of routine diagnostic work.
From "Ruijin" to the Grassroots: Deployed in 90+ Hospitals
The launch of RuiPath 2.0 builds on the success of RuiPath 1.0, released in February 2025 as China's first clinical-grade pathology model integrated into hospital production workflows.
Ning Guang, an academician of the Chinese Academy of Engineering and President of Ruijin Hospital, stated at the launch that the purpose of developing this pathology model is "not to build a 'high wall' of technology in top-tier hospitals, but to build a 'highway' to the grassroots".
This "highway" is now taking shape: Over 90 hospitals nationwide, including Chongqing University Three Gorges Hospital, Pu'er People's Hospital, Rui'an People's Hospital, and Shijiazhuang People's Hospital, have officially adopted the model as part of Huawei Cloud's "Industry AI Dream Factory". Leveraging this platform, hospitals need only less than 10% of the local data typically required for traditional training to develop their own customized pathology models through reinforcement learning, continuously improving accuracy via a "data flywheel" mechanism.
The release of RuiPath 2.0 marks a critical turning point for Chinese medical AI, moving from "SOTA in academic papers" to "productivity in hospital wards." It directly addresses the three major barriers to clinical adoption—accuracy, interpretability, and cost—by surpassing international benchmarks, providing transparent reasoning, and enabling deployment on standard PCs. When the diagnostic capabilities of a top-tier hospital are extended to county-level facilities via the cloud, AI transcends mere technological novelty to become a foundational element of healthcare equity.