OpenAI Launches “Sora” A Text-to-Video AI Model

OpenAI has unveiled its cutting-edge text-to-video AI model, Sora, as part of its ongoing 12-day “Ship-mas” product release series. Available now on Sora.com for ChatGPT subscribers in the US and many other countries, this revolutionary model generates videos directly from text prompts, enabling limitless creative possibilities.

The platform also introduced Sora Turbo, an upgraded version that offers advanced features like animating images, remixing videos, and creating videos from sequential text prompts.

Subscription Plans

Accessing Sora’s features requires a ChatGPT subscription. With a ChatGPT Plus plan priced at $20 per month, users can create up to 50 videos (1,000 credits) at 720p resolution with a duration of up to 5 seconds.

For $200 per month, the ChatGPT Pro plan provides unlimited video generations, 500 priority videos, 1080p resolution, 20-second video durations, watermark-free downloads, and simultaneous generation of up to five videos.

Key Features and Innovations

During a livestream event, OpenAI showcased Sora’s remarkable capabilities:

• Storyboard Tool: Allows users to generate videos using a sequence of text prompts.
• Animation from Images: Converts still images into animated videos.
• Remix Tool: Refines and tweaks AI-generated videos with text-based instructions.
• Blend Function: Combines two video scenes seamlessly using AI.

Additionally, Sora features a community-driven explore page where users can browse AI-generated videos. To ensure responsible use, OpenAI has implemented strict guidelines against uploading explicit, violent, or copyrighted content. Videos also include visible watermarks and metadata for transparency.

Addressing Concerns

OpenAI acknowledges challenges in content moderation. According to Sora’s product lead, Rohan Sahai, the company aims to balance safety with creative freedom. OpenAI encourages user feedback to refine its moderation system.

Although available in the US and many other countries, Sora’s rollout in Europe and the UK will take longer. For non-subscribers, the explore page remains accessible for viewing community-created content.