Google's Gemini AI Achieves 91.1% Accuracy in MedQA Benchmark, Extends into Medical Domain

May 2, 2024, 07:10 AM

Google has introduced a new AI model named Gemini, developed by DeepMind and Google Research, featuring three versions: Ultra, Pro, and Nano. The model, which has been extended into the medical domain as Med-Gemini, excels in a variety of medical applications, including answering medical questions and analyzing medical images and videos. Med-Gemini has achieved state-of-the-art performance on 10 out of 14 benchmarks, surpassing GPT-4 in all comparisons, and achieves 91.1% accuracy on the MedQA (USMLE) benchmark. It also features advanced reasoning, access to up-to-date medical knowledge, and understanding of complex multimodal data.