Veo 3.1, Google’s artificial intelligence now runs in 4K and focuses on vertical videos

One of the most significant innovations concerns the alignment with current trends in multimedia consumption, thanks to the introduction of native support for the 9:16 vertical format. If previously this option was limited to text prompts, it is now also extended to generations based on visual references, eliminating the need for manual cropping in post-production. This functionality finds a natural place within the Mountain View ecosystem, with the debut of Veo tools in the Gemini app and their direct integration into YouTube Shorts and YouTube Create, simplifying the workflow for creators targeting mobile platforms.

On the technical quality front, Veo 3.1 takes a significant step forward in the resolution of final content. Although the generation process still maintains 720p as a starting point, the new upscaling system now allows clips to be exported in 4K, surpassing the previous limit of 1080p. Google has also refined the rendering of the Full HD resolution to offer sharper and cleaner images, trying to bridge the gap with the promises of high fidelity made during the first presentations of the model. Although it is not yet a native generation in very high definition, the implementation of these visual improvement algorithms represents concrete progress towards quality standards closer to the needs of image professionals.