We're thrilled to unveil new updates and capabilities to help developers like you build the future with Google AI this year at Cloud Next. From our newest Gemini 2.5 thinking models, to new advancements in the Live API for real-time interaction, and Veo 2 becoming generally available for high quality video generation, here's a look at some of the exciting announcements this week for developers using the Gemini API in Google AI Studio.
We recently introduced Gemini 2.5 Pro, our most capable AI model, showcasing the power of thinking models that can reason before responding. Our most advanced coding model yet, Gemini 2.5 Pro excels at creating visually compelling web apps and developing agentic programming applications.
Just last week, Gemini 2.5 Pro became available for developers to build with the Gemini API in Google AI Studio and for enterprise customers, Vertex AI.
Link to Youtube Video (visible only when JS is disabled)
Building on this momentum, we're excited to share that Gemini 2.5 Flash is coming soon. This evolution of our popular workhorse model will maintain low latency and cost-efficiency while incorporating thinking capabilities.
This marks a significant step in our vision to make all Gemini models adaptively think. Building with Gemini 2.5 models unlocks a host of new use cases for applications enabling more capable agents, managing multi-agent systems, and accelerating code assistance and generative reasoning about entire code bases with a one million token input context window.
We're excited to announce that Veo 2 is now production ready in the Gemini API. Veo 2 is able to follow both simple and complex instructions, as well as simulate real-world physics in a wide range of visual styles. Veo 2 empowers developers to generate high-quality videos directly within their applications from both text and image prompts:
For example, Wolf Games is building a generative gaming platform that creates personalized interactive story games. Using Veo 2, they build dynamic cinematic experiences, benefiting from significantly enhanced video realism, motion accuracy, and camera control. Wolf Games say they slashed the iterations needed to get visuals right by over 60% and substantially reduced production time, bringing their creative vision closer faster.
Link to Youtube Video (visible only when JS is disabled)
Veo 2 is available today in the Gemini API in Google AI Studio:
Ready to build interactive applications with video generation? Dive into our documentation, prompt guide, and the getting started cookbook for Veo 2. Read more about Vertex AI’s enterprise-grade generative media across other modalities like speech and music.
Dynamic, real-time interactions are crucial for next-generation AI applications. The Live API for Gemini models is now in Preview, enabling developers to start building and testing more robust, scalable applications with significantly higher rate limits. Try the latest features now using the Gemini API in Google AI Studio and in Vertex AI.
The Live API enables developers to build applications and agents that process streaming audio, video and text with low latency, perfect for creating human-like conversations, participating in live meetings, or monitoring real-time situations.
Since its experimental launch in December, we've incorporated extensive developer feedback, adding highly requested features to the GA release:
Combined with powerful tool integrations (search, code execution, function calling), these features make the Live API ideal for using models like Gemini 2.0 Flash in highly interactive applications.
Ready to build real-time experiences? Dive into our documentation and check out the getting started cookbook for the Live API.
We are excited about the possibilities these updates unlock for the developer community. From more powerful thinking capabilities with Gemini 2.5, to real-time interactions via the Live API and video generation with Veo 2, we can't wait to see what you build next!
