Discontinued - verified 2026-09-30
The Sora app and web ended on 2026-04-26. OpenAI’s deprecations page schedules Sora 2 and the Videos API (
sora-2,sora-2-proand snapshots) for removal on 2026-09-24, with no recommended replacement. Wikipedia (secondary) states the API was discontinued on that date, ending the Sora brand; OpenAI’s own page only announces the schedule, so actual removal is reported rather than independently observed. Kept as history; for current video models see google-veo.
OpenAI Sora
by openai
Text-, image- and video-to-video generation model and social app (2024-2026).
What it was
A diffusion-transformer video model (space-time patches) first previewed in 2024. Sora 2 (announced late 2025) added native synchronised audio, better physics and a consumer app with cameos.
Why it ended
OpenAI gave no official reason; reports cite compute shortages, cost pressure and a shift to enterprise products. Wikipedia also reports that Disney’s USD 1 billion investment and 3-year character-licensing deal (Dec 2025) ended with the shutdown; treat these secondary details as unverified.
Consequences
- Anything built on the Videos API needed migration before 2026-09-24. Aggregators such as kie-dot-ai still advertise “Sora” in their docs; whether it works is unverified.
- Alternatives: google-veo, kling, seedance, minimax-hailuo-video, wan-2-dot-5, Runway (see runway).
- OpenAI image generation continues: gpt-image-2-5.
Sources
- https://developers.openai.com/api/docs/deprecations (accessed 2026-09-30)
- https://en.wikipedia.org/wiki/Sora_(text-to-video_model) (accessed 2026-09-30)