Google has launched Gemini Pro, a lightweight version of its flagship GenAI model family, for Google Cloud customers using Vertex AI. Gemini Pro is accessible in public preview in Vertex AI through the new Gemini Pro API, which supports 38 languages and regions and features like chat functionality and filtering. Gemini is a state-of-the-art multimodal model that allows developers to build their own applications. The Gemini Pro API accepts text as input and generates text as output, while the Gemini Pro Vision endpoint can process text and imagery. Gemini Pro can be customized to specific contexts and connected to external APIs for improved accuracy. Citation checking and grounding features help improve the quality of the model’s responses. Gemini Pro on Vertex AI is priced attractively, and Google is offering discounts and a free trial. Other new features in Vertex AI include conversational voice and chat agents, search features, and Automatic Side by Side for model evaluation. Google is also adding models from third parties and extending its indemnification policy.
