Google’s updated Veo model can make vertical videos from reference images with 4K upscaling

In a bold move for 2026, Google has unveiled the latest update to its Veo video AI, known as Veo 3.1, building on the impressive advancements made in 2025. This upgrade introduces exciting new features that allow users to create vertical videos using reference images, a capability dubbed Ingredients to Video. With this innovative approach, creators can input up to three images, which the AI will use to generate videos. This means users can animate characters, craft backgrounds, and select material textures more effectively. The upgraded model is designed to produce outputs that more closely align with the provided images, resulting in fewer unexpected changes. Additionally, the Veo 3.1 now allows for the generation of multiple video clips, offering users the flexibility to request alterations in settings or styles while maintaining the overall consistency of other elements. A significant enhancement is the focus on mobile-oriented video creation. The Ingredients to Video feature now supports a 9:16 aspect ratio, making it perfect for social media platforms such as Instagram, TikTok, and YouTube Shorts. This development reflects Google’s commitment to catering to the growing demand for mobile-first video content, following the introduction of vertical video support via text prompts last year.