With Veo 3.1, Google will let users add multiple reference images to guide the final output.
For developers, the Veo 3.1 model in the API is priced the same as the Veo 3 model
Photo Credit: Google
Google released the first major update to its artificial intelligence (AI) video generation model Veo 3 on Wednesday. Dubbed Veo 3.1, the updated model arrives less than five months after the release of Veo 3 and brings significant improvements to prompt adherence and granular control over the final output. Users can now add reference images to guide the video, and even place the first and final frame of the video to let the AI connect the dots. The model is currently not available on the Gemini app.
In a post on X (formerly known as Twitter), the official handle of Google DeepMind announced the release of Veo 3.1 video AI model. The company said that in the five months since Veo 3's launch in May, users have generated more than 275 million videos, highlighting its popularity. The new update, the tech giant claims, focused on user feedback and brings more artistic controls. Currently, Veo 3.1 is available via the Flow app and in the Gemini API for developers.
Broadly, there are three new features. The first is dubbed “Ingredients to Video,” which lets users upload multiple reference images while generating a video. The AI analyses the images and integrates them into the output. Google says this will allow users to generate videos that are closer to the creator's vision.
The second feature is “Frames to Video,” and with this, users can add a starting and ending image. The AI model will then generate a video that connects the two points. This feature is intended to help users get their desired narrative from the AI-generated videos with creative transitions.
Finally, the third feature is called Extend. Users can upload a clip, and Veo 3.1 continues the shot to build on top of it. Google says this mode will let users generate longer videos that last more than a minute. These videos are generated based on the final second of the uploaded clip to create continuity and help users when they want a longer establishing shot.
Notably, while the Flow app is available to users with the Google AI Pro and Google AI Ultra subscriptions, developers can access Veo 3.1 via the Gemini API. The pricing has been kept the same as Veo 3, with each second of the video generation being charged at $0.40 (roughly Rs. 35). Additionally, the Veo 3.1 Fast model will cost devs $0.15 (roughly Rs. 13) per second of generation.
Catch the latest from the Consumer Electronics Show on Gadgets 360, at our CES 2026 hub.