Sora

What Sora Was

Sora was OpenAI's video generation model. The original model appeared in February 2024, and Sora 2 later added more accurate physics, instructions that carried across several shots, and synchronized dialogue and sound effects. It turned written prompts into video with audio, and OpenAI released it alongside a social iOS app where people remixed each other's clips and placed themselves or friends into scenes through a characters feature. A developer API ran in parallel with the app.

What Happened

OpenAI discontinued the Sora web and app experiences on April 26, 2026. OpenAI announced that API access would end on September 24, 2026. The official help article covers both dates, along with content exports, data deletion and refunds: people could export what they created, associated data is permanently deleted once any final export window ends, and purchased ChatGPT and Sora credits can still be used for Codex.

As of October 2026, the old Sora address redirects to a notice stating that Sora is no longer available, with an option to export data. The article does not name a successor video product.

Alternatives to Sora

These tools from the directory cover the ground Sora covered, each with a different strength:

• Google Veo 3.1 suits creators who want dialogue and sound effects generated together with the picture, especially those who already pay for a Google AI plan.
• Runway Gen-4.5 suits filmmakers and developers who want cinematic text-to-video or image-to-video shots and a public API.
• Kling AI suits short narrative or serialized content that needs the same character to stay consistent across shots.
• Pika Art suits social creators who want quick, effects-driven clips from a prompt or an image, with a free tier to try first.
• Luma AI suits teams that want to chain generation steps into repeatable workflows, with commercial-use rights included.

What to Check When Switching

Audio comes first. Sora 2 produced sound together with the picture, and the listings above differ on this: Veo 3.1 and Kling AI generate audio, while Runway Gen-4.5 does not, so plan a separate audio step there. Anything built on the Sora API needs a new integration, and API terms and billing differ by tool. Prompts written for Sora will need rewriting, so test a few of your own on a free tier before committing. Anyone who relied on the characters feature should check which tools keep a person consistent between shots. Finally, confirm commercial-use terms for the plan you choose before publishing generated footage.