OpenAI has officially launched its hyperrealistic AI video generation software, Sora, 10 months after its initial preview in February 2024. The newly released Sora Turbo comes with significant upgrades and is accessible to ChatGPT Plus and Pro subscribers in the U. S. and many countries outside the EU and UK at sora. com. CEO Sam Altman announced the release through a YouTube livestream as part of the "12 Days of OpenAI" series. Sora allows users to create videos from text or images, with durations of 10 to 20 seconds and resolutions ranging from 480p to 1080p. It offers different aspect ratios and features a unique user interface, including a Storyboarding mode for generating linked clips with seamless transitions. ChatGPT Plus subscribers can create up to 50 videos monthly at 480p, while the Pro plan offers higher resolutions and unlimited generations at slower speeds.

OpenAI plans to offer tailored pricing options by early 2025. Tech reviewer Marques Brownlee, aka MKBHD, revealed the news on social media just before the official announcement. He shared examples of videos created with Sora, noting its impressive results but also its tendency to hallucinate details, such as nonsensical text and unnatural physics. OpenAI enforces strict guidelines to prevent generating real person likenesses and restricts violent or explicit content. Sora's release follows a protest leak on Hugging Face by beta testers unhappy with OpenAI's handling of the beta program. They criticized the company for relying on unpaid contributions while offering limited compensation through a competition. The AI video generation landscape is becoming increasingly competitive, with companies like Runway, Luma AI, and several Chinese firms launching similar tools. Although OpenAI gains recognition from its ChatGPT success, it faces competition from other platforms offering comparable or superior video quality, making Sora's success uncertain.