🔘 Google released Veo 3.1

The update is not revolutionary, but there are noticeable improvements in video generation control.

Main features:

• Control over narration and sound. The model now generates dialogues and ambient sounds better. This is promised to reduce the time needed for post-production.
• Flexible input data. Text, images, references, clips — can be combined and you can set style, start and end frames, continue scenes, and edit objects within the video.

Veo 3.1 is already available in Flow, Gemini API, and will soon appear in Vertex AI. Prices remain the same: 40 cents per second for the standard model and 15 for Fast. Videos are generated in 720p or 1080p, with durations up to 148 seconds when "stretching" scenes.