Google’s Med-Gemini AI: A Leap Beyond GPT-4 in Medical Innovation

Google unveiled its newest line of medical-focused artificial intelligence (AI) models. Although the tech giant has issued a pre-print version of its research paper that illustrates the capabilities and methodology of these AI models, dubbed Med-Gemini, they are not accessible for public usage.

According to Google, in benchmark testing, the AI models outperform the GPT-4 models. This specific AI model’s long-context capabilities, which enable it to process and analyze research papers and health records, are one of its standout qualities. The study work is available online at arXiv, an open-access repository for academic research, and is presently in the pre-print stage. 

Also Read: Gemini 1.5 Pro: Key Features, Price & How To Use This Next-Generation Model

In a post on X (formerly known as Twitter), Jeff Dean, Chief Scientist at Google DeepMind and Google Research, expressed his excitement about the potential of these models to improve patient and physician understanding of medical issues. I believe that one of the most significant application areas for AI will be in the healthcare industry.

Another post on X shows a hypothetical demonstration of how Med-Gemini works.

Gemini 1.0 and Gemini 1.5 LLM serve as the foundation for the Med-Gemini AI models. The Med-Gemini-S 1.0, Med-Gemini-M 1.0, Med-Gemini-L 1.0, and Med-Gemini-M 1.5 are the four versions in total. All of the models have multiple output formats, including text, image, and video. 

The company states that the models are enhanced by self-training, making them “more factually accurate, reliable, and nuanced” when displaying outcomes for intricate clinical reasoning tasks. The models are also integrated with web search.

The company further states that the AI model has been optimized for better performance during long-context processing. If the chatbot had better long-context processing, it would be able to respond with greater precision and accuracy even in situations where the questions are not correctly asked or when it needs to process lengthy medical record documents.

Read the official document released by Google here.

Google has released data showing that Med-Gemini AI models have performed better on text-based reasoning tasks in the GeneTuring dataset than OpenAI’s GPT-4 models. Also, Med-Gemini-L 1.0 outperformed its own earlier model, Med-PaLM 2, by 4.5 percent, with an accuracy score of 91.1 percent on MedQA (USMLE). 

Also Read: How to Use Gemini AI in Google Docs, Sheets, Slides, Gmail, and Drive?

Notably, the AI model is not available for public use or beta testing. The company will likely make more changes to the model before releasing it to the broader public.