Open- and closed-source LLMs in medical and engineering education

Back to news list

Source: Frontiers Medicine

Original: https://www.frontiersin.org/articles/10.3389/fmed.2025.1751813...

Published: 2026-01-13T00:00:00Z

The study compared the performance of open-source (DeepSeek, GLM-4, Kimi) and closed-source (GPT-4) models in training medical and engineering students. DeepSeek outperformed other models in all question types and achieved the highest accuracy rate. After applying rapid engineering techniques such as role-playing, knowledge generation, and brainstorming, the accuracy of the models improved significantly. DeepSeek exceeded 95% accuracy for all question types after training. Short-answer questions performed best with up to 97% accuracy among all four models. The findings highlight the potential of open source models to support medical and engineering education and highlight the importance of rapid engineering in problem solving.