This study evaluates how effectively Large Language Models (LLMs) can diagnose diseases based on patient symptoms, with DeepSeek-R1 achieving 76% accuracy in disease identification.

Key Findings:

• Compared diagnostic performance of DeepSeek-R1 and O3 Mini across chronic conditions
• Assessed accuracy at both specific disease and broader category levels
• Evaluated the reliability of model confidence scores
• Tested across multiple medical domains including Mental Health, Neurology, Oncology and Respiratory Disease

Business Impact: This research demonstrates the growing potential of LLMs as diagnostic support tools for healthcare professionals, potentially improving early detection, reducing misdiagnosis, and enhancing clinical decision-making at scale.

LLMs in Disease Diagnosis: A Comparative Study of DeepSeek-R1 and O3 Mini Across Chronic Health Conditions