At I/O 2025, Google stunned audiences with Veo 3, its advanced AI video generation model, and Flow AI, a creative filmmaking tool that transforms prompts and images into lifelike videos complete with dialogue and sound effects. The demos quickly went viral—but Google isn’t stopping there.
Five months later, the company is back with Veo 3.1, a major update that builds on the success of the original model. Announced Wednesday, the new version—first spotted in recent Gemini 3.0 leaks—delivers enhanced realism, richer audio, and greater narrative control. Flow, now powered by Veo 3.1, gains a suite of advanced editing features designed to streamline AI-powered filmmaking.
Where to Access Veo 3.1
Veo 3.1 will be available directly in the Flow and Gemini apps, and for developers via the Gemini API and Vertex AI. A faster variant, Veo 3.1 Fast, will also be offered across the same platforms for lower-latency generation.
What’s New in Veo 3.1 and Flow
According to Google’s blog post, Veo 3.1 offers stronger prompt adherence and improved audiovisual fidelity, particularly when converting still images into motion. The model supports both horizontal (16:9) and vertical (9:16) formats across text-to-video and image-to-video modes.
Flow also introduces several upgraded tools powered by Veo 3.1:
- Ingredients to Video: Combine multiple images—characters, objects, or styles—to craft complete video scenes.
Available in Flow, Vertex AI, Gemini API, and the Gemini app. - Frames to Video: Provide a starting and ending image, and Flow will generate the in-between footage to connect
them. Available in Flow, Vertex AI, and Gemini API. - Extend: Seamlessly lengthen videos using the last second of a clip as a visual and audio seed. Currently
available in Flow and Gemini API, with Vertex AI support coming soon.
Smarter Editing and Object Control
To make editing more flexible, Flow now includes Insert—a powerful feature that allows users to place any object or creature into a scene, whether realistic or fantastical. Flow automatically adapts shadows and lighting to maintain realism. The feature is available in Flow, Vertex AI, and the Gemini API.
The opposite is also possible: unwanted objects can be cleanly removed. Flow reconstructs the background to appear untouched, launching first on Vertex AI before expanding to Flow users. Additionally, Nano Banana image editing is already active in Flow for fine-tuning specific visual details.
With Veo 3.1, Google continues to push the boundaries of generative video technology—blurring the line between professional editing and creative expression powered by AI.
