OpenAI's Sora 2 is here: Not just for generating videos, but a social app that makes you the star

OpenAI has launched its new video generation model, Sora 2, which not only makes significant leaps in realism and physics simulation but also comes with a social app called ‘Sora’ that allows users to ‘insert’ themselves into any AI-generated scene. Let’s see how this technology will change the way we create and interact.

Imagine not just describing a movie scene in words, but actually becoming the protagonist of that scene. This sounds like something out of a science fiction novel, but with OpenAI’s official release of its latest flagship video generation model, Sora 2, on September 30, 2025, it’s quickly becoming a reality.

Sora 2 is not just an ordinary upgrade. It far surpasses previous systems in terms of simulating the physical world, realism, and controllability, and even adds the amazing ability to generate dialogue and sound effects simultaneously. Even more exciting is that OpenAI has also launched a social application of the same name, “Sora,” which is poised to subvert our imagination of content creation and social interaction.

This is not just video generation, it’s more like a “world simulator”?

In the past, AI video generation models, when faced with complex instructions, would sometimes produce… well, somewhat strange results, such as distorting the laws of physics to complete a task. But Sora 2 is a completely different story.

It has made significant progress in simulating the real physical world. For example, if you instruct a basketball player to shoot but miss, Sora 2 will faithfully show the ball bouncing off the backboard, rather than forcing the ball to “miraculously” pass through the hoop. This ability to accurately simulate “failure scenarios” is considered a key step towards becoming a truly useful world simulator.

This new model can now handle many complex instructions that were previously unimaginable. Want to see the perfect movements of an Olympic gymnast? Or the precise interaction of water spray and gravity during a backflip on a paddleboard? Or even, completing a triple axel in figure skating while holding a cat on your head? Sora 2 can do all of this.

Sound and picture, finally perfectly synchronized

One of the most important breakthroughs of Sora 2 is that it is no longer just a “video” generator, but a general “audiovisual” generation system.

It can directly generate videos from text prompts and “simultaneously” create background sound effects, ambient sounds, and even character dialogue that perfectly match the picture. This means that the lip movements of the characters in the video can be aligned with the voice, and the ambient sound effects will change with the dynamics of the picture. This completely changes the tedious process of AI videos requiring additional dubbing and post-production, providing a complete immersive experience.

Be an AI director: precise control of style and shots

The instruction understanding ability of Sora 2 has also been greatly improved, allowing users to precisely control the visual style of the video like a director. Whether you want a realistic, cinematic, or anime style, it can be presented in high quality.

You can even perform multi-shot narrative control, specifying the order, rhythm, and shot changes of the shots. Sora 2 will ensure the continuity of characters and environment between scene switches, effectively avoiding common problems in previous AI videos such as sudden changes in clothing, light jumps, or disappearing props.

New social App “Sora”: you are the protagonist

We’ve talked a lot about technology, but what really brings Sora 2 into the public eye is probably the simultaneously launched iOS social application “Sora”.

Its core feature is called “Cameos”. Sounds cool, right?

The operation is very simple: users record a short video in the App to verify their identity and capture their portrait and voice. After completion, you can accurately “implant” yourself or your friends’ appearance into any virtual scene generated by Sora. Want to dance in front of the Eiffel Tower, or play table tennis with a panda? Now, you can really be the protagonist of the video.

Born for creation, not addiction

OpenAI emphasizes that their original intention in designing this App was to “maximize creation, not consumption.” In other words, it is not to make you scroll your phone endlessly. To launch this powerful technology responsibly, they have also taken a number of measures:

• You are in charge of your feed: You can use natural language commands to tell the App what kind of content you want to see, giving you complete control over your news feed.
• Strengthen community connections: The App initially adopts an invitation system, encouraging you to use it with real friends. OpenAI believes that the “Cameos” feature will truly strengthen the connection between people.
• Protecting teenagers: The App has a default daily limit on the number of generated videos that teenage users can watch, and has stricter permissions for the “Cameos” feature. Parents can also perform more detailed parental controls through ChatGPT.
• Your portrait, your say: The most important point is that users have complete control over the use of their portraits. You can decide who can use your Cameo at any time, and you can revoke or delete any video containing your portrait at any time.

Want to try it out? How to get Sora 2

Are you excited? Currently, the Sora iOS app has been released for download in the United States and Canada, and there are plans to quickly expand to other countries. You can register in the App first, and you will receive a push notification when your account is activated.

After receiving the invitation, you can also use the functions of Sora 2 through the official website sora.com.

The good news is that Sora 2 will be available for free in the initial stage, with a very generous generation quota. In addition, if you are a ChatGPT Pro user, you can also use the experimental, higher-quality Sora 2 Pro model on the website. OpenAI also plans to open Sora 2 to developers in the form of an API in the future.

• Product official website: https://sora.chatgpt.com/
• Official introduction: https://openai.com/index/sora-2/

The launch of Sora 2 and the app of the same name is undoubtedly another milestone in the development of AI. It not only blurs the line between the creator and the creation, but also opens up new possibilities for personal digital expression and social interaction. What will happen next? Let’s wait and see.