OpenAI · Tools

Sora

OpenAI's text-to-video model.

At a glance

Pricing model
paid
Price
Pricing on request
Hosting region
Global
Vendor
OpenAI
Report data / suggest a correction

Profile

Sora is OpenAI's video-and-audio generation model — currently Sora 2 and Sora 2 Pro — which turns a text prompt or a reference image into a short video clip with synchronized sound. As of July 2026 it can only be used through OpenAI's developer Videos API, billed per second of generated video: OpenAI discontinued the consumer sora.com website and mobile app on April 26, 2026, and has announced that the API itself will shut down on September 24, 2026.

Who builds it

Sora is built by OpenAI, the same company behind ChatGPT, the GPT API, and DALL-E. The original Sora model was first previewed in 2024; OpenAI describes the current Sora 2 generation as its "flagship video and audio generation model." Access has narrowed sharply since launch: what began as a public-facing app and website is now, as of July 2026, an API-only product for developers, with an announced end date of its own.

Core features

  • Text-to-video and image-to-video — generates video clips from a written prompt or from a reference/input image.
  • Synchronized audio — Sora 2 produces video with matching sound rather than silent clips, a step up from the original Sora.
  • Two quality tiers — sora-2 is built for fast, flexible exploration, while sora-2-pro targets production-quality output with higher-resolution exports.
  • Standard video formats — Sora 2 outputs portrait (720×1280) or landscape (1280×720) video by default.
  • Video management, not just generation — the Videos API supports creating, extending, and editing videos, not only one-shot text-to-video requests.
  • Strict content policy — the API rejects real people (including public figures), copyrighted characters, and copyrighted music, and restricts output to content suitable for audiences under 18.

Pricing

Sora is billed per second of output video with no flat subscription, since the consumer app has been shut down. On the Standard tier, sora-2 costs $0.10/second at 720p; sora-2-pro costs $0.30/second at 720p, $0.50/second at 1024p, and $0.70/second at 1080p. A Batch tier cuts these rates roughly in half — down to $0.05/second for sora-2 and $0.15–$0.35/second for sora-2-pro, depending on resolution. There is no free tier, and OpenAI has confirmed the entire Sora API will be discontinued on September 24, 2026.

Who it's for

Sora is for developers who need to generate short AI video clips programmatically — not for casual consumers, since the consumer-facing app and website no longer exist and there is no walk-up way to use Sora without an OpenAI API account and billing set up. Given the confirmed September 2026 API shutdown, it currently suits short-term or experimental integrations more than long-term product bets; OpenAI has indicated Sora will continue afterward mainly as an internal research project rather than a public product. Teams evaluating Sora today should factor the shutdown date into any roadmap decision and have a migration plan for the underlying video-generation feature before the API goes dark.

Frequently asked questions

What is Sora?

Sora is OpenAI's flagship video-and-audio generation model — currently Sora 2 and Sora 2 Pro — that creates short video clips with synchronized sound from a text prompt or a reference image, accessed through OpenAI's Videos API.

Is Sora still available, and how do I access it today?

Only through OpenAI's developer API, and even that is temporary. OpenAI discontinued the consumer sora.com website and mobile app on April 26, 2026, and has announced that Sora API access (the Videos API, sora-2 and sora-2-pro) will also shut down on September 24, 2026.

How much does Sora cost?

Sora is billed per second of generated video. Standard-tier API pricing is $0.10/second for sora-2 (720p), and $0.30/$0.50/$0.70 per second for sora-2-pro at 720p/1024p/1080p; a Batch tier cuts these rates roughly in half.

What's the difference between Sora 2 and Sora 2 Pro?

Sora 2 is built for fast, flexible exploration, while Sora 2 Pro is aimed at production-quality output with higher-resolution exports up to 1080p, at roughly three to seven times the per-second price of standard Sora 2.

Can I generate videos of real people or copyrighted characters with Sora?

No. OpenAI's content policy for the Sora API explicitly rejects real people (including public figures), copyrighted characters, and copyrighted music, and restricts generations to content suitable for audiences under 18.

What happens to videos I already created with Sora?

OpenAI says users can export content they created in Sora, and that data tied to Sora usage will be permanently deleted once the discontinuation — and any final export window — has passed.