At this time, we’re saying Gemini 2.0 — our most succesful AI mannequin but, designed for the agentic period. Gemini 2.0 has new capabilities, like multimodal output with native picture era and audio output, and native use of instruments together with Google Search and Maps.
We’re releasing an experimental model of Gemini 2.0 Flash, our workhorse mannequin with low latency and enhanced efficiency. Builders can begin constructing with this mannequin within the Gemini API by way of Google AI Studio and Vertex AI. And Gemini and Gemini Superior customers globally can check out a chat optimized model of Gemini 2.0 by deciding on it within the mannequin dropdown on desktop.
We’re additionally utilizing Gemini 2.0 in new analysis prototypes, together with Undertaking Astra, which explores the long run capabilities of a common AI assistant; Undertaking Mariner, an early prototype able to taking actions in Chrome as an experimental extension; and Jules, an experimental AI-powered code agent. We proceed to prioritize security and duty with these initiatives, which is why we’re taking an exploratory and gradual method to improvement, together with working with trusted testers.
This week, we began testing Gemini 2.0 in AI Overviews in Search, and early subsequent yr, we’ll increase Gemini 2.0 to extra Google merchandise.
