Sora is an AI video generation service developed by OpenAI. From a simple text prompt, it automatically generates high-quality videos up to 20 seconds long at 1080p, with an emphasis on physically consistent motion. Sora 2, released in September 2025, added support for videos with synchronized audio, and the service played a major role in popularizing AI video generation. However, OpenAI announced the service’s discontinuation in March 2026: the app and web version ended on April 26, 2026, and the API is scheduled to shut down on September 24, 2026. This article looks at what Sora offered and what you can still do with it today.
Key Features
- Text-to-video generation: Generate scene videos from natural-language instructions like “a cat walking through Tokyo streets at dusk.” With careful prompting, users could also control camera work and visual style
- Physically consistent visuals: From its debut, Sora drew attention for producing footage with realistic physics — object motion, light reflections, and the behavior of water and smoke
- Audio support in Sora 2: The September 2025 release of Sora 2 enabled videos with synchronized sound effects and dialogue
- Up to 20 seconds at 1080p: Short-form output, but with resolution and length sufficient for social media posts and concept footage
- API integration: Developers could embed video generation into their own apps and workflows via the API (available until September 24, 2026)
Pricing
| Offering | Status | Notes |
|---|---|---|
| App / Web | Discontinued (April 26, 2026) | Previously available to ChatGPT Plus / Pro subscribers via a credit system |
| API | Scheduled to end September 24, 2026 | Usage-based billing per second of generated video (rates vary by model and resolution) |
Pricing information is current as of August 2026. For exact API rates, check the official pricing page. Given the shutdown date, new production adoption is not recommended.
Pros & Cons
✅ Pros
- Create high-quality videos from text alone, with no video production expertise
- Well regarded for natural footage with physical consistency
- Sora 2 added audio support, producing ready-to-use material
- Built by OpenAI, the maker of ChatGPT, sharing the same account and ecosystem
⚠️ Cons
- The service’s discontinuation is confirmed, and the app / web version is already unavailable
- The API also ends on September 24, 2026, making it unsuitable for new workflow integrations
- Videos were limited to 20 seconds, making it a poor fit for long-form content
- The service became a focal point of copyright and deepfake controversies, requiring care over generated content
Comparison with Similar Services
| Comparison | Sora | Runway | Google Veo | Kling AI |
|---|---|---|---|---|
| Provider | OpenAI | Runway | Kuaishou | |
| Availability | Ending (API until September 2026) | Active | Active | Active |
| Video with audio | Supported (Sora 2) | Partial | Supported | Supported |
| Strengths | Physical consistency, prompt adherence | Integrated editing tools for filmmakers | Integration with Gemini / Flow | Longer, high-quality human motion |
| API | Yes (ending) | Yes | Yes | Yes |
Who Is It For
- Developers who want to keep using the API until its shutdown (September 24, 2026) while preparing to migrate existing video generation workflows
- People following the history and technology trends of AI video generation who want to understand Sora’s impact
- Users with previously generated Sora content who need to download and back it up
- For anyone starting with AI video generation now, we recommend considering actively maintained services such as Runway or Google Veo instead
Summary
Sora demonstrated to the world that text prompts could produce physically consistent, high-quality video, and it drove the AI video generation boom. However, following a shift in OpenAI’s business priorities, the app and web version ended on April 26, 2026, and the API will shut down on September 24, 2026. If you have generated content, download it soon; for new video generation needs, consider actively maintained alternatives. Sora itself is said to continue not as a video product but as a research project focused on world models.