Google Unveils Veo 3.1 in Response to OpenAI: How to Access the New AI Video Model

Earlier this month, OpenAI unveiled Sora 2 with great fanfare. Touted as the video equivalent of the “advent of GPT-3.5,” it was paired with a new video-sharing platform inspired by TikTok. It took Google just two weeks to respond.

In a blog post published on Wednesday, October 15, 2025, the company revealed the Veo 3.1 model along with new advancements for Flow. These updates “provide users with even more creative control”. Let’s dive in!

Google Launches Veo 3.1: What’s New

Veo 3.1 follows the previous Veo 3 model. Users will experience the following enhancements:

• Audio Generation: Google notes they are “integrating audio into existing features to help you create the perfect scene”.
• Enhanced Visual Rendering: Textures, lighting, and movements are more natural.
• Narrative Consistency: Transitions between shots are smoother and maintain visual continuity better.

Flow Adds New Features

Along with the new model, Flow, the creation platform launched by Google five months ago, is updated with these new options:

• Ingredients to Video: Flow allows the use of several reference images to define the style, characters, or objects in a scene.
• Images to Video: The tool connects a starting and an ending image to generate a continuous sequence.
• Extend: Flow can extend a clip from its last image to create longer, more cohesive shots.
• Insert: This new feature allows adding new elements into a scene while maintaining existing light and shadows.
• Remove: Flow automatically reconstructs the setting when an object or character is removed, yielding a natural result.

How to Access Veo 3.1 in Flow

To access the new features, visit the Flow platform, available at this link. Initially launched only in the United States, it is now available in most countries worldwide, including France. However, it remains exclusive to subscribers of Google AI Pro and Google AI Ultra plans. Free users seem to have very limited access to the Veo 3.1 model, although the exact amount of credits available is not specified.

Once on the platform, you can enter your request from the lower interface. Three options are offered: “Text to Video”, “Images to Video”, and the new “Ingredients to Video”. Two options from Veo 3 are also displayed: “Quality” (takes longer to generate but with a more accurate output) and “Fast”, which produces results in reduced time.

Developers can also take advantage of the new features through the Gemini API, which allows integrating Veo 3.1 functions into their applications, while businesses can access it via Vertex AI.