Update (October 2026): Veo 3.1 now comes in Standard, Fast and Lite variants. On Google's Gemini API, the Veo 3.1 preview models are scheduled to shut down on October 22, 2026, with Gemini Omni Flash as the recommended replacement there; generally available Veo models continue on Google's enterprise platform. Read Gemini Omni Flash vs Veo 3.1 for details.
Google released Veo 3.1 as an update to its Veo 3 video generation model, refining several capabilities rather than introducing an entirely new architecture. The changes focus on richer audio, better prompt adherence, and new creative controls like scene extension and reference images. This guide breaks down what's actually new in Veo 3.1, how it compares to Veo 3, and where you can access it.
What Is Veo 3.1?
Veo 3.1 is Google DeepMind's updated video generation model, building on Veo 3 with targeted improvements rather than a full architectural revision. It generates short video clips typically around 8 seconds with natively generated audio, meaning sound is produced alongside the visuals rather than added separately afterward.
The model is available through the Gemini API, Google AI Studio, Vertex AI, the Gemini app, and Flow, Google's AI-powered video editor. A lightweight version, Veo 3.1 Fast, and a lower-cost version, Veo 3.1 Lite, are also available for different production needs.
What's New in Veo 3.1
Richer Native Audio
Veo 3.1 generates more natural conversational audio and better-synchronized sound effects than the original Veo 3, including improved lip sync for subjects who are speaking.
Reference Images
Users can now upload multiple reference images up to three, including people, characters, or products to help guide the generated video toward a specific look or maintain visual consistency with existing assets.
Scene Extension
Veo 3.1 can extend a previously generated clip by creating new segments that continue from the final second of the prior video, maintaining visual continuity. This makes it possible to build longer sequences, sometimes over a minute, from a series of connected generations.
Frame-Specific Generation
The model supports generating a video by specifying a first and last frame, giving more direct control over how a scene begins and ends rather than relying entirely on prompt description.
Portrait Video Support
Veo 3.1 supports both landscape (16:9) and portrait (9:16) video orientations, making it easier to produce content built specifically for mobile and vertical social platforms without cropping afterward.
Improved Prompt Adherence
Google reports that Veo 3.1 handles complex, multi-element prompts — combining subject, action, environment, camera direction, and style — more reliably than the original Veo 3, with fewer frame-to-frame inconsistencies across a clip.
Video Upscaling
A separate Veo upscaling capability, rolled out alongside Veo 3.1 on Vertex AI, allows existing lower-resolution videos to be enhanced up to 1080p or 4K, regardless of whether the original footage came from Veo, another AI model, or a traditional camera.
Veo 3.1 vs. Veo 3: What Actually Changed
Veo 3.1 is best understood as a refinement of Veo 3 rather than a new generation of the model. The core architecture is the same; the update focuses on:
More natural and better-synchronized audio
Additional creative controls (reference images, frame-specific generation, scene extension)
Better consistency for subjects and scenes across longer or more complex prompts
New format support (portrait video)
If you already had access to Veo 3 through a Google product, the update to 3.1 typically applied automatically — the version generally isn't something you select manually within consumer-facing tools.
Where to Access Veo 3.1
Veo 3.1 and Veo 3.1 Fast are available through:
Gemini API (Google AI Studio and Vertex AI) for developers building applications
Gemini app for conversational, prompt-based generation
Flow, Google's dedicated AI video editor, for more hands-on creative control
Access and usage limits vary by product and plan, so checking Google's current documentation for the specific tool you're using is the most reliable way to confirm what's included.
Using Veo 3.1 Alongside Other AI Video Models
Veo 3.1 is a strong option for creators who want native audio and tighter prompt control, but it's one of several capable video models available today. Rather than committing to a single provider, many creators prefer platforms that give access to multiple AI video models so they can compare results for a given prompt.
TextVidai works this way — you can generate AI videos from text using an AI video generator that supports multiple models, rather than being limited to one. If you want to see finished examples across different styles before starting, the AI-generated video examples on the showcase page are a useful reference point.
Frequently Asked Questions About Veo 3.1
What is Veo 3.1?
Veo 3.1 is Google DeepMind's updated video generation model, building on Veo 3 with improvements to audio quality, prompt adherence, and creative controls like reference images and scene extension.
What's new in Veo 3.1 compared to Veo 3?
Veo 3.1 adds richer native audio, reference image support, scene extension, frame-specific generation, portrait video support, and improved consistency for complex prompts.
How long are Veo 3.1 video clips?
Veo 3.1 generates clips of around 8 seconds by default, though scene extension can be used to create longer sequences by connecting multiple generations.
Does Veo 3.1 generate audio?
Yes. Veo 3.1 generates audio natively alongside video, including conversational dialogue and synchronized sound effects, rather than requiring a separate audio-generation step.
Where can I use Veo 3.1?
Veo 3.1 is available through the Gemini API, Google AI Studio, Vertex AI, the Gemini app, and Flow, Google's AI video editor.
Is Veo 3.1 the only good AI video model available?
No. Veo 3.1 is one of several capable models available today, and platforms like TextVidai give access to multiple AI video models rather than relying on a single provider.
Conclusion
Veo 3.1 refines what Veo 3 already did well, adding richer native audio, reference images, scene extension, and portrait video support without changing the underlying model architecture. For creators, the most practical improvements are better consistency across longer or more complex prompts and more control over how a video begins and ends. If you want to try Veo 3.1 alongside other AI video models rather than committing to one provider, you can use TextVidai to generate and compare results.