OpenAI Launches Sora 2: More Realistic AI Videos — Now with Sound

OpenAI isn’t wasting any time. Less than a year after releasing its first video-generation model, the company is back with Sora 2, a major upgrade that delivers far more realistic videos — and now comes with sound.

But that’s not all. Alongside the model, OpenAI is also rolling out Sora, a brand-new iOS app designed for creating, sharing, and remixing AI-generated videos. Think of it as a TikTok-style social network, but with content that’s 100% powered by AI.

Sora 2: More Natural Videos with Built-in Audio

With this second version, OpenAI has fixed many of the early limitations. Sora 2 can now handle:

• More detailed scenes
• More natural movements
• Logical interactions between characters and objects
• Better scene-to-scene consistency

The biggest upgrade, however, is sound. Sora 2 can now generate audio alongside video, meaning you can request a complete scene with voices, background noise, sound effects, and even music — all in one go. In theory, this removes the need to edit audio separately in a video editor.

Cameos: Put Yourself in the Scene

One of the most talked-about features in Sora 2 is cameos. Users can record a short clip of themselves (video + voice) to create a digital version that can appear in any AI-generated scene.

Whether it’s a sci-fi set, a sports highlight, or something totally absurd, your AI clone can take part — no need to film yourself in those environments. In short, it’s like creating safe, personalized deepfakes of yourself.

OpenAI even allows collaborative cameos: you can give friends permission to use your cameo so you can all appear together in the same generated scene.

On the safety side, OpenAI emphasizes:

• Cameos are opt-in and can be removed at any time
• Public figures are blocked by default to prevent abuse

Sora App: An AI-Powered Social Platform

Alongside the model, OpenAI has launched Sora, a dedicated mobile app currently available only on iOS in the U.S. and Canada.

The app works like a social video platform:

• Users can create or remix AI-generated videos
• Videos are shared in a personalized feed similar to TikTok or Instagram Reels
• People can interact with others’ creations, remix them, or add their cameo (with permission)

Unlike traditional social media, the focus is on creative interaction, not passive scrolling.

To address common criticisms of social networks, OpenAI has included several safeguards:

• Algorithm controls via natural language prompts
• Parental controls powered by ChatGPT
• Well-being checks and limits for teen accounts

Limited Access and Progressive Rollout

As with many of its launches, OpenAI is rolling out Sora 2 gradually:

• Only a limited number of users currently have access via the iOS app (invite-only)
• Users can sign up for notifications when access opens up
• The model is free at launch, with “generous” usage limits
• Paid credits may be introduced later if demand is high

Meanwhile:

• ChatGPT Pro subscribers can test Sora 2 Pro, a more advanced version, via sora.com and soon in the app
• An API is also on the way, enabling developers to integrate Sora 2 into other tools and services

The Next Step in AI-Generated Media

Sora 2 shows just how fast AI-generated video is evolving — moving from silent, rough animations to full multimedia experiences with sound and user personalization. With the launch of the Sora app, OpenAI is positioning itself not just as a toolmaker, but as a platform for AI-first creativity and social interaction.

The big question now: will users embrace an AI-only content ecosystem the same way they did TikTok or Instagram?

❓ FAQ: OpenAI Sora 2

What is OpenAI Sora 2?
Sora 2 is OpenAI’s latest AI video generation model. It creates more realistic videos with smooth motion, coherent scenes, and—for the first time—built-in sound.

Does Sora 2 generate sound?
Yes. Unlike the first version, Sora 2 can generate audio such as voices, ambient sounds, effects, or music along with the video.

What are Sora cameos?
Cameos let you upload a short clip of yourself (image + voice) to create a digital version of you that can appear in AI-generated videos.

Is Sora available on all devices?
Currently, the Sora app is only available on iOS in the United States and Canada. Android support has not been announced yet.

How can I access Sora 2?
Access is limited and invite-based for now. Users can sign up through the Sora iOS app or use the Pro version on sora.com if they have a ChatGPT Pro subscription.

Will Sora 2 have an API?
Yes. OpenAI confirmed that an API for Sora 2 will be available soon, allowing developers to integrate the model into apps and services.