Succession decided by six companies at once
This entry aggregates models from six makers, so succession is not one question but six, and none of them belongs to the company selling the subscription. Nothing on the page says how a maker's retirement would be passed on, and one listed model already has a published end date. From 2 source pages, read 2026-09-12; no maker pays to be listed.
| What the page is asked | What invideo AI publishes | What that leaves open |
|---|---|---|
| Is a successor named | Nothing on the page names one | For any of the six makers represented |
| Who decides succession | Each maker, on its own schedule | Six schedules, none of them this company's |
| Is a dated line still listed | One listed model has a published end date | Recorded from the maker's own announcement |
| How would a retirement arrive | Not described anywhere on the page | Presumably as a line leaving the menu |
Inclusion rule. Read from the app's own published model list. Dates beside those models come from whoever makes them and are recorded on the calendar, not invented here. Order. The absence first, then who actually holds the decision.
1Aggregation multiplies the question, not the answer
A single-model entry has one unannounced future. This one has six, and the company a customer pays has authority over none of them. That is the structural point, and it holds regardless of how well the app is run.
It also cuts the other way. When one maker retires a line, five others remain on the menu, so the customer keeps a working product and loses a particular look. Whether that is a consolation depends entirely on whether a series was built on the line that went.
2A listed model with a published ending
One model on this list carries a date published by its maker. Both statements are current: the app names the model, and the maker says the interface behind it closes. The register keeps them side by side without resolving the tension.
Resolving it would mean predicting what the app does next, which is exactly the kind of inference this register refuses. A menu page is a marketing document and menus lag; the useful habit is checking the maker.
3What a subscriber can actually watch
Not this page. A menu can change without a notice, so the informative surface is each maker's own changelog, and there are six of them with no shared format.
That is the cost of aggregation stated plainly, and it is why the register keeps model pages apart from entry pages. The entry says what is purchasable; the maker says what is ending, and only one of those two ever carries a date.
- Sora API24 September 2026discontinued
- Named replacement modelNo replacement named (as of 2026-09-12)none in the announcement
4Production that outlasts a model
Model end dates touch “invideo AI”; SceneMixer's guides describe how its pipeline treats the models it runs:
Doubao Seed 2.1 Pro is the family’s full-size tier at 100% of the base text rate, twice the default. Same three jobs — reading, writing, storyboards — with more headroom on long, crowded episodes: it holds the storyboard’s strict format more reliably and keeps track of many characters and locations in one pass.
SceneMixer, Doubao Seed 2.1 Pro (read 2026-10-13)
Dubbing replaces the audio of a finished clip; the picture was generated for a different line, so mouth shapes, pauses and shot length all belong to the original language. Native-language generation gives the video model the line in the target language before it renders the shot, so the performance, the pacing and the cut points are built for that language.
SceneMixer, Native-Language Dialogue vs. Dubbing in AI Video (2026): Why the Cast Should Speak the Line (read 2026-10-13)
5Sources
Read from the model list on invideo.io, 2026-09-22. The whole entry is on invideo AI, and the column it sits in is on Named replacement. Same column, other entries: Adobe Firefly, Microsoft Foundry.