Trying New Sora 2 From OpenAI and Sharing The Invite
Sora 2 has been released by OpenAI — a next-gen text-to-video model and app focused on more realistic physics, stronger scene/control fidelity, and built-in sound. For now it’s invite-only (please check comments you can find one), with access rolling out gradually to select creators and waitlist users; broad public timing isn’t announced yet.
Here are some samples we have generated using Sora2
- Sora2 generates photorealistic clips up to ~1 minute with better world/physics modeling, stronger multi-character consistency, wider stylistic range, and built-in sound (speech, SFX, ambience).
- “Upload yourself” / Cameos Record a short verification clip and Sora can insert your likeness and voice into scenes; you control who can use your cameo and can revoke it anytime.
- Editing & launch constraints Rollout is conservative on rights and safety (tighter moderation; more controls for rightsholders and for personal likeness). Policies are evolving after industry feedback.
- Access & where to use it iOS app at launch (U.S./Canada), invite-only to start; after you’re invited you can also use the web app at sora.com. Android is planned later.
- Context & growth The app crossed 1M downloads in <5 days despite invites and regional limits; OpenAI says moderation and creator controls are being expanded