Google DeepMind Enhances Veo 3.1: Introducing Professional Controls and Native Vertical Video

January 13, 2026 — Google DeepMind has announced a major feature update for its Veo 3.1 video generation model. The update centers on the "Ingredients to Video" toolset, designed to provide creators with professional-grade precision when transforming reference images into cinematic motion.

Native Vertical Output and 4K Upscaling

Veo 3.1 is now optimized for the mobile-first era with the introduction of native 9:16 vertical support. This allows creators to generate high-quality, full-screen content for platforms like YouTube Shorts directly, avoiding the resolution loss typically caused by cropping horizontal footage.

To meet the standards of high-end production, Google has also implemented advanced 4K upscaling. This feature enhances textures and fine details, making AI-generated clips suitable for professional workflows and large-screen displays.

Advanced Control: The "Ingredients" System

The core of the update is a significant breakthrough in identity and stylistic consistency. The enhanced "Ingredients to Video" framework now offers:

• Triple-Image Referencing: Users can provide up to three reference images (such as a character, an environment, and a visual style) to guide a single generation. This ensures the AI maintains the character’s likeness across different scenes.
• First and Last Frame Control: A powerful new capability allows creators to set the exact starting and ending points of a video. Veo 3.1 then generates a seamless, cinematic transition between the two images, preserving visual logic throughout the clip.
• Scene Extension: To facilitate longer narratives, the model can now extend videos by generating new segments based on the final frames of a previous clip, ensuring continuity in both visuals and motion.

Cinematic Intelligence and Integrated Audio

Veo 3.1 now interprets professional cinematography commands—such as "dolly-in," "pan," or "cinematic lighting"—with greater accuracy. Additionally, the model generates synchronized audio, including ambient sounds, sound effects (SFX), and experimental lip-syncing for characters, creating a more immersive and cohesive viewing experience.

Availability and Security

These new Veo 3.1 features are being integrated across the Google ecosystem, including the Gemini app, the Flow filmmaking platform, and for enterprise users via the Gemini API and Vertex AI.

To ensure transparency, all content generated by Veo 3.1 is embedded with SynthID digital watermarks. Furthermore, users can now verify the origin of a video by uploading it to the Gemini app to check if it was produced using Google’s AI tools.