Veo 3.1 Puts Google in Direct Competition with OpenAI’s Sora 2

Google DeepMind has announced Veo 3.1, the latest version of its generative video model, alongside expanded capabilities in Flow, its AI filmmaking tool. The update introduces audio generation, new editing options, and greater control over video creation.

“Five months ago, we introduced Flow, our AI filmmaking tool powered by Veo, and have been inspired by the creativity it has sparked with over 275 million videos generated,” said Google DeepMind in a statement. 

Veo 3.1 offers better narrative control and realism, building on the Veo 3 model with stronger prompt adherence and improved audiovisual quality when converting images into videos. The model is now integrated with Flow and available through the Gemini API, Vertex AI, and the Gemini app.

The new release brings audio support to existing Flow tools such as Ingredients to Video, Frames to Video, and Extend, allowing creators to produce videos with synchronised sound. 

The company said the features are still in development and will be refined further based on user feedback.

With Ingredients to Video, users can control the appearance of their scenes using multiple reference images. Frames to Video lets them generate smooth transitions between two static images, while Extend allows the creation of longer, continuous shots that build on previous clips.

Flow also adds new editing tools. The Insert feature enables users to add new elements, such as characters or objects, into scenes, while an upcoming Remove function will allow unwanted objects to be deleted seamlessly. “Flow now handles complex details like shadows and scene lighting, making additions and removals appear natural,” the company said.

The Veo 3.1 launch comes just days after OpenAI released Sora 2, its latest video and audio generation model, along with a social iOS app called Sora that lets users create, remix, and insert themselves into generated videos.