Google introduces Veo 3.1 with audio and advanced editing in Flow

Google has introduced Veo 3.1, the new version of its video generation model that, according to the company, represents the current state of the art in this field. The model brings three main improvements over its predecessor Veo 3: richer audio, greater narrative control, and improved realism in texture capture. Veo 3.1 also improves instruction adherence and audiovisual quality in image-to-video conversion.

Veo 3.1 capabilities are implemented in Flow, Google's audiovisual creation tool that enables video generation through artificial intelligence. Since its launch five months ago, the platform has generated over 275 million videos. The updates respond to user requests for greater artistic control and audio support across all functionalities.

Flow now incorporates improvements in two directions. On one hand, it adds generated audio to existing functions that previously only produced video without sound. The "Ingredients to Video" option allows using multiple reference images to control characters, objects, and visual style, generating scenes according to the user's vision. "Frames to Video" creates transitions between a starting and ending image, producing a video that connects both smoothly. The "Extend" function enables creating longer videos, up to a minute or more, that continue the action from the original clip based on its final second. These three functionalities, which already existed in Flow, now include enriched audio generated by Veo 3.1.

On the other hand, Flow adds completely new editing capabilities. The "Insert" function allows adding elements to any scene, from realistic details to fantastical creatures, managing aspects like shadows and lighting to integrate elements naturally. Soon a function will be available to remove objects or characters from scenes, automatically reconstructing the background and environment to make it appear the element was never there.

Veo 3.1 is available through the Gemini API for developers, Vertex AI for enterprise customers, and the Gemini app. Flow's new capabilities are accessible through both the Gemini API and Vertex AI. Google notes these functions are experimental and undergoing continuous improvement.

Key points

• Google has launched Veo 3.1, its new video generation model that the company describes as current state of the art.
• Veo 3.1 brings three main improvements: richer audio, greater narrative control, and improved realism in texture capture.
• Flow, Google's audiovisual creation tool, has generated over 275 million videos in five months since its launch.
• Flow adds generated audio to three existing functions that previously only produced video: "Ingredients to Video," "Frames to Video," and "Extend."
• New "Insert" function allows adding elements to scenes with automatic shadow and lighting management.
• A function to remove objects or characters by automatically reconstructing the background will be available soon.
• Veo 3.1 is available in the Gemini API, Vertex AI, and Gemini app.