Sora: 184 days' notice, no replacement. Veo previews: Omni named
OpenAI gave 184 days between its Sora notice and shutdown and named no replacement. Google's Veo 3.1 preview table names gemini-omni-1.1-flash.

OpenAI announced the Sora API shutdown on March 24, 2026 and shut it down on September 24, 2026, which is 184 days, and its deprecation table lists no replacement. Google's table for the Veo 3.1 preview ids gives an October 22, 2026 shutdown date and names gemini-omni-1.1-flash as the replacement for all three. If you are choosing what to migrate to, the second table answers the question and the first does not.
The two tables, side by side
Everything in this table is read from the vendors' own pages on 2026-10-08. The Google page lists a shutdown date and a replacement for each id but, in the part this post read, no announcement date, so the notice length for Veo is left blank rather than estimated.
| Item | OpenAI (Sora) | Google (Veo 3.1 previews) |
|---|---|---|
| Announced | March 24, 2026 | Not stated in the table |
| Shutdown | September 24, 2026 | October 22, 2026 |
| Ids affected | sora-2, sora-2-pro, sora-2-2025-10-06, sora-2-2025-12-08, sora-2-pro-2025-10-06, Videos API | veo-3.1-generate-preview, veo-3.1-fast-generate-preview, veo-3.1-lite-generate-preview |
| Replacement named | None listed | gemini-omni-1.1-flash |
| Notice | 184 days | Not stated |
What a missing replacement means in practice
A blank replacement column does not mean there is no good answer; it means the vendor did not pick one for you. Teams that used Sora had to choose from other vendors' catalogs, and that choice depends on what you used: text-to-video, image-to-video, longer clips, or a specific look. Sume's docs list the models it can run and their limits, and the model to test first for a short, general clip is gemini-omni-flash-1.1: 3 to 10 seconds, 360p to 4K, 16:9 or 9:16, with native audio. For longer clips, Sume's Seedance 2.5 and Wan 3.0 rows accept up to 30 seconds.
What the Google table gets you
Google's replacement is a different family of model, not a new Veo: its video page says to use Gemini Omni Flash as the default model for video generation, and keeps Veo 3.1 for features such as video extension and frame-specific generation. If your code only needs text-to-video or image-to-video, the named replacement covers it. If it uses Veo's extension, the Omni docs describe their own extension, which appends 3 to 10 seconds up to 40 seconds in total.
The id differs between Google and Sume. Google's replacement id is gemini-omni-1.1-flash; Sume's catalog id is gemini-omni-flash-1.1, bare and without a provider prefix. Copy the Sume id from GET /v1/videos/models, not from Google's table.
Counting the days you have left
OpenAI's window is closed: September 24 has passed, and the Sora API no longer accepts calls. Google's is open for 14 more days from today, October 8, to October 22. A 184-day notice is long enough to plan a quarter, and teams that waited for a named replacement got none. The Google notice is the opposite problem: the replacement is named, but you must test it against your own prompts quickly. A reasonable plan is one day to inventory ids, three days to run 20 to 30 real prompts through the replacement, and the rest of the window for a staged cutover.
A decision rule
Use the vendor's replacement column when it exists, then verify the replacement against the target platform's own catalog before you rely on it. When the column is empty, as with Sora, write down the three features your calls used, check each against a candidate model's supported_durations, supported_resolutions and supported_input_references, and test one clip per candidate. Dates and tables change, so re-read both vendor pages on the day you act.
Sources
Related posts
More in Comparisons
- Speech-to-speech vs cascade: does 42% lower latency matter for video?
A newsletter reports speech-to-speech models run 42% lower latency than cascades. Why that helps live agents, not a narrated video, and what to measure.
- Speech-to-text per hour: Sume $0.60 batch vs MAI streaming $0.54
Sume STT is batch at $0.01 per audio minute, $0.60 an hour. MAI-Transcribe-2-Streaming lists $0.54 an hour through year end. Different products; the math.
- sume/auto or a pinned image model: what the response shows
With sume/auto, job.model stays sume/auto, the model list omits it and the family is never named. A pinned id echoes back as requested.
- Sume Format vs pasting the same prompt each time: what changes
A Format stores the house style once, then each API run sends only inputs. How the instruction is composed, what a run returns, and the spend cap.
Written by Sume