OpenAI Unveils Sora 2 Model and Launches Video-Sharing Social App

OpenAI has introduced Sora 2, the next-generation text-to-video model, along with a new social media app that enables users to create and share AI-generated videos. The launch places OpenAI in direct competition with Google’s Veo-3, which gained significant traction among users earlier this year.

Advancements in Realism and Controllability

OpenAI believes that Sora 2 can perform more precise calculations with regard to complex physical interactions than its predecessor. To illustrate, when a basketball shot is missed, the ball rebounds naturally rather than showing up in unrealistic positions.

The model also improves controllability, enabling users to follow multi-shot instructions while maintaining consistency in the visual environment. It supports cinematic, anime, and realistic video styles, while also adding background audio such as soundscapes, speech, and effects.

Features of Sora 2

Sora 2 introduces the ability to insert people or real-world objects directly into AI-generated videos. While OpenAI acknowledges that the model is “far from perfect”, the company views it as evidence that scaling neural networks with video data can enhance realism.

From Creation to Cameos: OpenAI Introduces the Sora App

To complement the model, OpenAI has launched a dedicated Sora app on iOS. The app allows users to create and remix videos, explore content through a customisable feed, and join scenes via a cameo feature. After a one-time video and audio recording, users can virtually insert themselves into generated clips.

OpenAI says it has built a recommendation system that can be guided through natural language, giving users more control over their feed.By default, the app prioritises content from followed creators and highlights videos considered likely to inspire new creations. Industry Context The introduction of Sora 2 and its companion app comes at a time when text-to-video models are becoming central to AI innovation. Google’s Veo-3 has already drawn significant attention, and OpenAI’s move underscores the growing competition to capture user engagement in AI-driven video creation.