Google is gearing up for a major announcement at I/O 2025, set for May 20. At this annual developer conference, the tech giant will debut its latest generative models, including both the new Imagen 4 and the enhanced Imagen 4 Ultra, alongside Veo 3 for advanced video production.
Early identifiers such as ‘veo-3.0-generate-preview’ and ‘imagen-4.0-ultra-generate-exp-05-20’ hint at a phased rollout reminiscent of previous Google Labs projects. These labels suggest a tiered approach that caters to varied needs in creative, commercial, and research settings.
Building on the foundations of Imagen 3 and Veo 2, these upcoming models are expected to offer better image fidelity, improved sequence coherence, and seamless multimodal integration. Recent integrations into tools like NotebookLM and Gemini further illustrate Google’s vision for a smooth transition from text prompts to visuals and video content.
I/O has long been a platform where Google showcases its high-profile innovations, and this year, the new releases could be central to the company’s 2025 strategy in generative media. Whether these models will soon be available to everyone or remain exclusive to select Labs users is still unclear, but the announcement reaffirms Google’s leadership amid tough competition from peers like OpenAI, Anthropic, and xAI.
