Google introduces a new video generation model as OpenAI shutters Sora app

Google’s not calling it quits on AI video, even as OpenAI shutters its Sora AI video app. The company announced a new Veo model today, aimed at developers looking to offer video generation in their own products.

Google introduced its new Veo 3.1 Lite model in a blog post today. The model is meant to balance “practical utility with professional capabilities,” the post says, costing developers as little as five cents per second for 720p generated video, including audio.

You won’t be able to generate videos with Veo 3.1 Lite in the Gemini app; it’s only available for paying Google AI Studio users. For those users, 720p video costs five cents per second and 1080p video costs eight cents per second. There are options to make landscape video clips in 16:9, or portrait in 9:16. The new model can’t generate 4K imagery.

There are options to create clips that are four, six, or eight seconds long — not suitable for trying to make movies, but useful for whipping up gen-AI animations to add to website or app interfaces.

Google’s also announced that its Veo 3.1 Fast model is coming down in price starting on April 7. Starting next month, that model can kick out 720p video for 10 cents per second (down five cents), 1080p video for 15 cents per second (down three cents), or 4K video for 35 cents per second (also a five-cent reduction).

Veo 3.1 Fast is rolling out for paid Google AI Studio users right now. You can check out the full developer documentation for the new model here.