Google launches Veo 3.1 video model with improved audio, vertical format & scene extension

Google has announced Veo 3.1 and Veo 3.1 Fast, the latest versions of its AI video generation model available through the Google Gemini app, Google Flow, Google AI Studio, and Vertex AI. Just a couple of weeks after OpenAI introduced Sora 2, Veo 3.1 arrives with key improvements in realism, fidelity, and prompt adherence, delivering higher quality videos with improved native audio that features natural dialogue and synchronized sound effects. The update also adds support for both landscape and portrait 16:9 formats, addressing the growing demand for vertical video.

Developers can now guide the model using up to three reference images to maintain consistent characters or objects across multiple clips, along with a new scene extension feature that enables longer videos by connecting new clips to the last frame of the previous one, preserving visual continuity (previously limited to 30-sec). The model also shows a deeper understanding of cinematic styles, giving developers greater creative control over tone and composition.

Veo 3.1 powers Google Flow filmmaking tool, which now supports generated audio for its “Ingredients to Video,” “Frames to Video,” and “Extend” features. Developers can upload images as start or end points, add custom audio, or generate transitions between scenes.

Both Veo 3.1 and Veo 3.1 Fast are available through the Gemini app, Google AI Studio, and Vertex AI at the same cost as Veo 3, with the Fast variant offering quicker and more affordable generations.