Google is giving its AI video lineup a big usability boost on Vertex AI, with a new Veo 3.1 Lite model for cheaper, high‑volume generation and a dedicated upscaling tool that can sharpen existing footage all the way up to 4K.

Veo 3.1 Lite is basically the “workhorse” tier of Google’s Veo 3.1 family: it’s built for teams that want to pump out lots of videos without burning through budget, while still getting professional‑looking results. It supports both text‑to‑video and image‑to‑video, handles common social and web aspect ratios like 16:9 and 9:16, and can render in 720p or 1080p, making it a natural fit for Shorts, Reels, and quick marketing clips. Google positions Lite as its most cost‑effective Veo option on Vertex AI, with third‑party cost guides pegging it at around 0.05 dollars per second of video, which is less than half the effective cost of Veo 3.1 Fast for similar workloads.

The broader Veo 3.1 lineup now has three tiers: Veo 3.1 at the top for premium, final‑cut‑ready visuals; Veo 3.1 Fast when you need speed with strong quality; and Veo 3.1 Lite for budget‑sensitive or large‑scale projects. All three can generate native audio along with video, so instead of stitching sound in a separate tool, you can get synchronized visuals and sound out of a single API call or workflow inside Vertex AI. For developers, that means you can mix and match tiers—ideate cheaply in Lite, move to Fast or full Veo 3.1 only when you lock in on a concept.

The new Veo upscaling capability solves a different but very familiar problem: “We already have footage, we just wish it looked sharper.” In private preview on Vertex AI, it lets you take low‑resolution clips—whether they came from Veo, another AI model, or a regular camera—and boost them up to 1080p or 4K using Google’s super‑resolution algorithms. Under the hood, Veo still synthesizes content at 1080p, then the upscaler adds detail and clarity, which is especially helpful for brands that need high‑quality assets for larger screens or premium ad inventory without re‑shooting or re‑generating everything from scratch.

For businesses that are already in the Google Cloud ecosystem, both Veo 3.1 Lite and the upscaling feature plug straight into Vertex AI APIs and Media Studio, so you can wire them into existing pipelines rather than standing up new infrastructure. A typical workflow could look like this: generate draft cuts at scale with Veo 3.1 Lite, run final picks through Veo 3.1 or Fast if you need higher‑end visuals, then upscale finished assets to 4K before sending them to your ad platform or content management system. That combination of tiers and upscaling makes Google’s video stack less about a single “hero” model and more about giving teams a toolkit for everything from scrappy social tests to polished campaigns.