Research

Source-grounded clinical answer engine (Zörgm Pro) versus general-purpose large language models on the NEET-PG 2025 examination

A comparative benchmark of a source-grounded clinical answer engine against four general-purpose large language models on 154 single-best-answer questions from the 2025 National Eligibility cum Entrance Test for Postgraduate (NEET-PG) examination. Zörgm Pro achieved the highest score of the five systems evaluated (98.5%) and was the only system to provide a cited source for every answer.

Arathy Varghese and Akhil Das · Laennec AI Limited

Preprint, for educational and research purposes. The authors are affiliated with Laennec AI Limited, the developer of Zörgm Pro; see the study’s declarations for competing interests.