Vertical Videos, Visual Consistency & 4K: Google Boosts Veo 3.1 Capabilities!

Google is rolling out a significant update to its Veo 3.1 video AI model, focusing on enhancements to the “Ingredients to Video” feature along with new export capabilities suited for mobile and professional formats.

Enhanced Expressiveness and Consistency with Ingredients to Video

The Ingredients to Video feature, which allows video generation from up to three reference images, has received major optimizations. According to Ricky Wong, Lead Product Manager at Google DeepMind, this update makes videos “more expressive and creative, even with simple prompts”.

Improvements have been made in several key areas of video generation:

• Character consistency: Veo 3.1 now maintains a character’s appearance consistently across different scenes and environments, making it easier to create complete narratives with the same protagonist.
• Background and object consistency: Users can reuse objects, settings, or textures from one scene to another while maintaining their visual integrity.
• Integration of disparate elements: The model can now coherently combine characters, objects, textures, and stylized backgrounds in a single clip.
• Enhanced dialogue and narration: The generated videos feature more dynamic movements and livelier storytelling.

Google recommends using the new Gemini 3 Pro Image model (also known as “Nano Banana Pro”) to create reference images for generating videos with Veo 3.1.

Native Vertical Format and High-Resolution Upscaling

The update introduces native support for the vertical format (9:16) in Ingredients to Video for the first time. This feature, previously available by request through text prompts, now allows the creation of content ready for publishing on TikTok, YouTube Shorts, and other short video platforms without manual cropping or quality loss.

Regarding image quality, Veo 3.1 now offers upscaling to 4K resolution, up from a maximum of 1080p previously. Google also notes that “the generation in 1080p has been enhanced to produce sharper and cleaner videos”. However, this 4K upscaling capability is separate from the native 4K generation that Google had announced for Veo in 2024, which has yet to be rolled out to the general public.

AI Content Verification with SynthID and Availability

In an effort to maintain transparency, especially as AI-generated deepfakes become more common, Google is incorporating its digital watermark, SynthID, into all videos generated by its tools. Since December, the Gemini app has allowed users to upload a video and verify if it was generated using Google’s AI technologies, adding to the already available image verification tools.

Furthermore, the new features of Veo 3.1 are being gradually deployed across the entire Google ecosystem:

• For creators and the general public: Enhancements to Ingredients to Video and the portrait mode are arriving in the Gemini app starting today, including through AI Pro subscriptions. YouTube Shorts and the YouTube Create app are also integrating these capabilities.
• For professionals and businesses: Flow, the Gemini API, Vertex AI, and Google Vids are receiving the improved features of Ingredients to Video and support for the vertical format. The 1080p and 4K upscaling is available on Flow, the API, and Vertex AI.

In Google Vids (available for Google Workspace), the Ingredients to Video feature with support for the vertical format has been fully rolled out since January 13, 2026, accessible notably to accounts including Business Starter, Standard and Plus, Enterprise, Nonprofits, and Education Plus.