OpenAI has unveiled Sora 2, its latest AI-powered video and audio generation model, alongside a dedicated Sora iOS app. The new model delivers improved realism, physics-aware interactions, synchronized audio, and greater creative control, building on the foundation of the original Sora launched in early 2024.
From Sora 1 to Sora 2
The first Sora was considered a “GPT-1 moment” for video, enabling object permanence and scaled training for basic world simulation. Sora 2 is described as a step closer to a “GPT-3.5 moment,” advancing the ability to model complex physics, realistic failures, and multi-scene continuity.
Key Features of Sora 2
- Physics-Aware Video: Models buoyancy, rigidity, collisions, and motion realistically.
- Realistic Failures: Errors resemble natural mistakes instead of distorted visuals.
- Controllability: Can follow multi-shot instructions and maintain consistent states across scenes.
- Visual Styles: Supports realistic, cinematic, and anime-style outputs.
- Integrated Audio: Generates synchronized speech, ambient sound, and sound effects.
- Personalization: Users can insert themselves or objects into AI-generated environments using short recordings.
The new invite-only iOS app serves as the main platform for Sora 2. Users can create, browse, and remix AI-generated videos; share content through a personalized feed; and use Cameos, which allow placing themselves or friends into scenes after a one-time video/audio recording.
OpenAI reports that internal testing showed Cameos encouraged new forms of social interaction, extending digital expression.
OpenAI is emphasizing safety with Feed Controls driven by natural language, not optimized for endless scrolling; Teen Protections including stricter cameo permissions, time limits, and parental controls; Consent & Control tools for revoking cameo access or removing videos; and Moderation via human reviewers and AI systems.
The company also confirmed no ad-based business model; optional paid tiers may only be introduced to balance compute demand.
Availability
- The iOS app launches first in the U.S. and Canada, with plans for global expansion.
- Users can sign up in-app for activation notifications.
- Web platform at sora.com mirrors app features.
- Pro Access: ChatGPT Pro users prioritized for early access.
- API Launch planned for developers.
- Legacy Support: Sora 1 Turbo remains available, with all prior creations preserved.
The Sora Team shared its view on the progress of AI video generation and its long-term implications: “Video models are improving rapidly. General-purpose world simulators and robotic agents will reshape society and accelerate human progress. Sora 2 marks important progress toward this vision. Consistent with our mission, we aim to ensure these models benefit humanity as they develop. We believe Sora will bring creativity, joy, and new ways of connection worldwide.”
