AI video models on imaginary.app

Models you can use today

Veo 3.1 is made by Google. We call the Google Cloud models directly, so you don't need a Google Cloud account. Veo 3.1 and Veo 3.1 Fast are generally available on Google Cloud; Veo 3.1 Lite is in Preview. Each variant has its own page with full credit tables, specs and real examples:

Starting from a photo instead of a prompt? The Veo 3.1 image-to-video guide covers uploads, first-and-last-frame and costs.

What an 8-second clip costs on each variant

Credits cost $0.21–$0.25 each, depending on the pack or plan. Shorter clips cost less; the model pages list every length.

To price a whole project, use the Veo 3.1 cost calculator. For how these numbers compare with Google Cloud and Flow (we cost more per clip than Google's list price), see each model page.

How to choose a model

• Still working out the prompt? Use Lite. It is the cheapest per clip, so iterating costs little.
• Making a finished social clip, ad cut-down or b-roll? Start with Fast, the generator's default. On Fast and Quality, 1080p costs the same as 720p.
• Hero shot, close-up dialogue, or a client deliverable? Use Quality. Our generator notes that the high-quality mode is more likely to generate sound.
• Need 4K? Fast or Quality. Google documents 4K for Quality; its docs are inconsistent about Fast in 4K, though our app offers it.
• Need clips longer than 8 seconds? No variant does that in one generation; chain clips with first-and-last-frame.

Our Lite vs Fast vs Quality comparison goes through the trade-offs in more depth, and the prompting guide applies to all three.

Shown in the app, but not offered

The generator also shows these models. None of them can be selected, none has a price, and you cannot generate with them on imaginary.app today.

Looking at other tools instead? Our comparisons and alternatives pages cover models we don't offer, with sourced specs.

What changed at Google in 2026

• June 30, 2026: Google retired Veo 3. Veo 3.1 replaced it; see Veo 3.1 vs Veo 3.
• October 22, 2026: the Veo 3.1 preview models on the Gemini API shut down. We use the Google Cloud models, so this didn't affect us. More in the Gemini API shutdown guide.
• December 5, 2026: Google Flow retires Veo 3.1 Fast and Quality (Lite stays). Both remain available here; see the Flow retirement guide.
• No Veo 4. Google's newer line is Gemini Omni, which we don't offer; see Veo 3.1 vs Gemini Omni Flash.

For developers

The same three variants are available through our REST API with one veo_ API key: send veo3-lite, veo3-fast or veo3-quality as the model to POST /api/generate. Credits and refunds work the same as on the web. See the API documentation.