SNUH Healthcare AI Research Institute

📢 Announcement

Stay updated with the latest major developments from the Healthcare AI Research Institute (HARI) at Seoul National University Hospital!

December 12, 2025

• 🧠 Model Release: A new medical reasoning model hari-q2.5-thinking is now available!
  • Our model achieved 89.2% accuracy on the KMLE and and 88.36% on the USMLE QA benchmark.

June 12, 2025

• 🧠 Model Release: Our first medical reasoning model hari-q3 achieved 84.1% accuracy on the KMLE and is now available!

May 24, 2025

• ✨ New Dataset: KorMedLawQA is now available!
  • We are thrilled to release KorMedLawQA, a new dataset featuring high-quality multiple-choice questions based on South Korean medical law. Each question is accompanied by detailed legal reasoning, designed to support the development of specialized medical LLMs for the Korean healthcare environment, assist in KMLE preparation, and serve as a valuable benchmark.

May 1, 2025

• 🧠 Model Release: hari-q2.5 achieving 84.6% accuracy on the KMLE is now available!
  • This advanced medical LLM is designed to support clinical decision-making and medical education. Stay tuned for more updates!

March 21, 2025

• 📊 Dataset Release: ClinicalQA dataset is now available!
  • We have released a dataset of over 1,000 high-quality question-answer pairs covering diverse clinical scenarios to enhance medical AI development and benchmarking.

🤗 Join us in shaping the future of AI-driven healthcare!

🌍 Vision & Mission

🔹 Vision: Shaping a sustainable and healthy future through pioneering AI research.
🔹 Mission:

• Develop cutting-edge AI technologies for clinical and operational excellence.
• Foster collaborations between clinicians, AI researchers, and industry leaders.
• Lead global healthcare AI commercialization and regulatory frameworks.
• Train next-generation AI-driven medical professionals.

🏗 Research Platform

• AI-Powered Medical Data Research Platforms: SUPREME, SNUHUB, DeView, VitalDB, Korea Health Data Platform
• Supercomputing AI Infrastructure: NVIDIA B200, H100, A100 GPUs, Quantum Computing Initiative.

💡 Key Initiatives

🔬 AI Development for Clinical Application

• AI-assisted Clinical note summarization, medical imaging diagnostics, real-time patient monitoring, and more!
• AI-powered electronic health records (EHR) automation.

📡 Global AI-Driven Healthcare Data Platform

• Secure cross-institutional medical data-driven AI research collaboration.
• Responsible AI frameworks for safe AI adoption in healthcare.

🎓 AI Training & Education

• Basic AI Education Program: Designed for physicians, researchers, and medical students, covering fundamental AI concepts, medical data analytics, and ethical AI usage.
• Advanced AI Research Program: Focused on AI model development, deep learning in healthcare, and clinical AI validation, aimed at senior researchers and specialists.

🤝 Collaborate with Us!

HARI is open to partnerships with:

• AI research institutions and academic collaborators.
• Healthcare technology companies and startups.
• Government agencies shaping AI policy in healthcare.

For inquiries and collaboration, please contact us at:
📧 hhoon@snu.ac.kr
🌐 Seoul National University Hospital

🚀 Together, we are shaping the future of AI-driven healthcare! 🤗