Google DeepMind rolled out Veo 3.1, an upgrade that turns static images into animated clips on mobile devices. The new version adds native 9:16 output, letting creators publish vertical videos directly to YouTube Shorts without cropping. Upscaling options now reach 1080p and 4K, delivering broadcast‑ready sharpness. The feature ships across the Gemini app, Flow, Google Vids, the Gemini API and Vertex AI for seamless integration.

Beyond aspect ratio, Veo 3.1 improves character and background consistency, so a protagonist retains its look across scene changes. Dialogue richness increases, giving clips a more natural narrative flow. Users can reuse objects or textures, blending them with stylized backdrops to produce cohesive stories. The Gemini app’s Nano Banana Pro model helps generate high‑quality ingredient images for the pipeline.

Developers can now embed Veo 3.1 into enterprise workflows via Vertex AI or the Gemini API, accessing 1080p and 4K rendering for marketing assets or internal training videos. Google DeepMind also integrates an invisible SynthID watermark, enabling easy verification of AI‑generated content. The rollout gives creators immediate tools to produce polished, mobile‑first video without external editing software.