Veo 3.1 preview API shuts down on October 22: three decisions for teams using it for AI dramas

Who this is for, and what it helps you decide

Creators and small teams who call the Veo 3.1 preview models directly through the Gemini API to make AI drama shots, motion-comic scenes or promotional clips, and anyone using a third-party tool who is not sure whether it relies on these models. It helps you decide which shots to finish today, which model to move to, and whether to re-test cost and visual consistency.

This covers only these preview model IDs in the Gemini API. We did not check whether Google's consumer products or third-party tools are affected.

What the official documentation says

The Gemini API deprecations page lists veo-3.1-generate-preview and veo-3.1-fast-generate-preview (released October 15, 2025) and veo-3.1-lite-generate-preview (released March 31, 2026) with a shutdown date of October 22, 2026, and recommends gemini-omni-1.1-flash as the replacement. The Veo page, the video generation overview and the pricing page all say these three preview models are "deprecated and will be shut down on October 22, 2026", and advise migrating to Gemini Omni Flash or the generally available models on the Gemini Enterprise Agent Platform "to avoid service interruptions".

Omni's own preview model, gemini-omni-flash-preview, is listed for shutdown on the same day, also with gemini-omni-1.1-flash as the replacement.

How firm is the date? The deprecations page notes that its dates are "the earliest possible dates" on which a model might be retired and that Google will give advance notice of the exact date, while the Veo and pricing pages say the models "will be shut down on October 22". No page gives a time of day or a time zone. The latest entry on the release notes page when we read it was dated October 13, and we did not see an entry specifically announcing the Veo 3.1 preview deprecation, so we cannot confirm when Google first published the date.

Some background: when Google announced in June that Veo 2.0 and 3.0 would shut down on June 30, it advised users to move to the Veo 3.1 preview model IDs. Teams that followed that advice now have to migrate again.

The two replacement routes

Route one is Gemini Omni Flash (gemini-omni-1.1-flash). Its documentation shows it is called through the Interactions API, not the generateContent API that Veo uses, so this means a code change, not just a new model name. Listed capabilities include text-to-video, image references, first-and-last-frame transitions, subject references, extending a video by a 3–10 second continuation at the end, and multi-turn conversational editing. It outputs 9:16 or 16:9 at 360p, 720p (default), 1080p or 4K; 1080p and 4K are upscaled.

Limits listed in the Omni documentation include: no uploading of images containing certain recognisable people; users in the European Economic Area, Switzerland and the UK cannot edit or extend videos they upload; uploaded videos for editing must be 10 seconds or shorter; no system instructions, temperature or negative-prompt parameters (negatives can go in the ordinary prompt); and no provisioned throughput. Both the Omni and Veo documentation say English is fully supported while other languages "have not been evaluated" and results may vary.

Route two is to stay on Veo. Google's video overview says to use Veo 3.1 when you need scene extension, last-frame control or integration with existing pipelines. Once the Gemini API previews shut down, the documentation points to the generally available veo-3.1-generate-001 and veo-3.1-fast-generate-001 on the Gemini Enterprise Agent Platform. That platform's model versions page lists their retirement date as "November 17, 2026 or later" and says retirement timelines may be extended but will not be moved earlier. We did not read that platform's pricing or sign-up requirements.

How the prices compare (official list prices, not tested by us)

On the Gemini API pricing page, Veo 3.1 preview is billed per second: Standard costs $0.40 per second at 720p or 1080p and $0.60 at 4K; Fast costs $0.10 at 720p, $0.12 at 1080p and $0.30 at 4K; Lite costs $0.05 at 720p and $0.08 at 1080p, with no 4K.

Gemini Omni Flash is billed by tokens: $1.50 per million input tokens and $17.50 per million video output tokens, with output pricing including thinking tokens. The page says 720p video is counted at 5,792 tokens per second, which it puts at roughly $0.10 per second. It is not available on the free tier.

On Google's own estimate, Omni at 720p costs about the same per second as Veo 3.1 Fast, about a quarter of Standard, and about twice Lite. Actual bills also depend on input, thinking tokens and retries; we have not measured them.

What it may mean for creators (editorial inference, not tested by us)

Today, search your code, configuration and automation scripts for these three model IDs. If the shutdown goes ahead on schedule, calls to them can be expected to fail, and queued batch jobs could stop all at once.

If you have unfinished sequences, consider generating and downloading those shots before the shutdown. The Veo documentation says generated videos are kept on the server for only 2 days, and extension works only on Veo-generated videos created or referenced in the last 2 days. Whether existing clips can be continued seamlessly on a new model is something to test yourself.

For a serialised drama or motion comic, a model change may alter characters' appearance and the visual style. Test with the same character reference images and a few key shots before switching a whole season, so one series does not change look halfway through.

Teams prompting in Chinese or other languages should note that both sets of official documentation promise full support only for English. If you use a third-party AI video tool, ask the vendor whether it depends on these preview models and when it will switch.