Performance of Large Language Models on Diagnostic Radiology Board–Style Questions: A Comparative Evaluation of GPT-4o, Perplexity AI, and OpenEvidence

Key Points

• The study aims to evaluate the performance of emerging language models in diagnostic radiology scenarios.
• Comparative evaluation of GPT-4o, Perplexity AI, and OpenEvidence on diagnostic questions.
• Utilized board-style questions relevant to diagnostic radiology to assess accuracy and reliability.
• Emerging LLMs like Perplexity AI and OpenEvidence demonstrated higher diagnostic reliability than traditional models.
• Performance metrics indicate improved accuracy in interpreting radiology questions.

Abstract