Sora (OpenAI)

Sora was OpenAI's AI video generator. You described a scene in text or with an image and it produced a short video, and you could edit it, for example by extending it or replacing part of it. A later version added video with sound and a social app that let people put themselves into videos.

OpenAI has since discontinued Sora, and new videos can no longer be created.

How the service was wound down

The dates come from OpenAI's help article "What to know about the Sora discontinuation." Press reports cited usage that did not grow enough to justify the operating cost. Check the official help article for how and until when existing videos can be exported.

Alternatives people are considering

• Google Veo: available from the Gemini app and Google's video creation tools, including generation with sound
• Runway: rich editing features aimed at video producers
• Kling: a video generation AI from China's Kuaishou
• Adobe Firefly: training data managed with commercial use in mind, connected to editing software such as Premiere Pro

All of these change their features and prices quickly. A production workflow that depends on a single service can stop working overnight, as Sora showed. Keep prompts, source assets and approval records on your side so you can switch services.

Cautions for business use of video generation

Do not generate the faces or voices of real people, or copyrighted characters and works. Even for your own ads, a person should check portrait rights and copyright. Never strip watermarks or metadata from generated video to pass it off as real footage; to guard against deepfakes, record internally that the content is AI-generated. Physically impossible motion and garbled text still occur, so a person should review and edit the final output.

Still images are covered in our image generation AI entry. We help companies introduce generative AI and write usage guidelines through our AI & DX services.