Google DeepMind has unveiled Veo 3.1, the latest iteration of its video generation model. The company frames the release around three core attributes: consistency, creativity, and control. The update is designed to produce clips that feel more natural and dynamic rather than stiff or artificial.
A notable new capability in Veo 3.1 is support for vertical video generation. This directly targets mobile-first and short-form video formats, which have become dominant in social platforms. According to DeepMind, the model now generates engaging clips that fit these vertical formats while retaining quality.
Because this analysis relies on a single source, the claims represent Google DeepMind's own description of the update. Independent testing or third-party evaluation would be needed to verify how consistent, creative, and controllable the outputs actually are. Still, the emphasis on vertical support suggests a clear shift toward practical, platform-ready video creation.