At the main event of Google I/O 2025, Google introduced its most advanced AI image and video generation tools, Imagen 4 and Veo 3. These new AI models can now not only create photos and video clips from text or image prompts, but automatic and accurate audio can also be generated in the video through Veo 3.

Veo 3: Now AI will create a realistic video and audio

According to Google, Veo 3 is its most advanced video generation model to date. It can do things like better movement and motion handling, stability between scenes, and better interaction with the environment. Now users can generate a few seconds of video clip by just providing text or an image, in which audio will also be automatically added. Veo 3 is currently available in beta on the Gemini app and Flow (Google's new AI filmmaking platform) for users of the Google AI Ultra plan. Enterprise users can access it through Vertex AI.

Veo 3 Now, continuity can be maintained in the scene by uploading an image of a person, object, or style. Apart from this, camera movements like pan, zoom, and rotate can now be decided by prompt. Now the video can be extended beyond its current frame. It is now possible to add or remove an object in a frame. Also, light and shadow are automatically adjusted.

Image 4: Now cleaner images
Google has also launched Imagen 4, which is a new AI image generator. Apart from high-quality image support up to 2K resolution, it has support for output in many styles from photo-realistic to illustrative with texture detailing, such as clothes, reflection, fur, etc.

PC Social media