Gemini 2.5 Pro Vectara new 7.0% - is that good for work docs

https://www.magcloud.com/user/william-roberts80

As of March 2026, the landscape for evaluating Large Language Models has shifted away from general chatbot performance toward hyper-specific metrics like hallucination rates on enterprise data. When I first saw the data point regarding Gemini 2