Google has announced significant updates to its AI tools at the Cloud Next 2025 conference. One of the key highlights is the wide availability of its Veo 2 video-generation model, which is now production-ready and accessible through the Google AI Studio’s Gemini API. Additionally, the company has teased the upcoming release of Gemini 2.5 Flash, a model designed for simple tasks coming soon.
Veo 2 stands out as a versatile video-generation tool capable of understanding both simple and complex instructions. It can simulate real-world physics and operate in various visual styles. Users can interact with Veo 2 using text or image prompts, allowing for a unique blend of creativity in generating videos.
Currently, Veo 2 produces videos at a resolution of 720p and 24 frames per second, although each video is limited to a maximum length of eight seconds. For enterprises, the service costs $0.35 per second of video. Google has provided examples of Veo 2’s capabilities, showcasing its features through partnerships like Wolf Games.
Alongside Veo 2, the Gemini 2.5 Flash model has been introduced, aimed at enhancing the existing Gemini 2.5 family of AI models that emerged last month. The company asserts that this new variant will uphold low latency and cost-effectiveness while enhancing thinking capabilities. In addition to these developments, Google has launched a preview of the Live API for Gemini, aimed at facilitating real-time interactions.
This tool allows developers to create applications that efficiently process streaming audio, video, and text. More announcements are expected from Google as Cloud Next 2025 continues.