OpenAI officially launched the Sora video generation AI model in its 12-day "ship-mas" product release series. The model is now online at Sora.com and is available to ChatGPT subscribers in the United States and most countries. The Sora Turbo version has many innovative features, including storyboards, style templates, editing, looping, remixing, mixing, etc. It supports the generation of videos up to 1080p resolution and up to 20 seconds long, and provides multiple ratio options. Users can create with text, images and videos, and use their own materials to expand, mix and generate new content. New interfaces and tools, such as storyboarding tools, also enhance the user experience.
During its 12-day "ship-mas" product launch series, OpenAI officially launched the highly anticipated Sora video generation AI model. This innovative technology is now available on Sora.com and is available to ChatGPT subscribers in the United States and most countries.
Sora Turbo introduces a series of exciting and innovative features:
Storyboard: Users can precisely control video storyboards through text, pictures or existing videos.
Style templates: Provides a variety of preset styles to meet different creative needs.
Re-Cut: The generated video can be extended forward and backward.
Loop: Create a seamless loop video from beginning to end.
Remix: Easily modify the generated video without rewriting the prompt words.
Blend: Achieve seamless connection between two videos.
Sora Turbo supports the generation of videos up to 1080p resolution and up to 20 seconds long, and provides a variety of aspect ratio options such as landscape, portrait and square. Users can create with text, images and videos, and can use their own materials to expand, mix and generate new content.
Sora also provides a new interface and tools, including storyboard tools, to help users create more easily with text, images and videos.
Well-known technology reviewer MKBHD has already had a preliminary experience with this product and unveiled a bit of mystery for users. Currently, Sora Turbo excels in color aesthetics, style presentation, and abstract art. However, there is still room for improvement in aspects such as physical simulation, motion capture, and semantic understanding. In fact, there is not a big gap with Runway, Keling, etc.
Usage details
Platform: Currently only supports the web side (sora.com)
Video generation:
Plus users ($20/month): 50 generation credits per month
Pro users ($200/month): 500 fast generation quota + unlimited slow generation quota per month
Video specifications: supports horizontal and vertical screens, providing three ratios: 16:9, 1:1, and 9:16
Maximum duration: 20 seconds (single generation)
Concurrent generation: Pro users can generate up to 5 videos at the same time
In terms of product design, OpenAI demonstrates impressive community capabilities. The platform provides an exploration page where users can browse AI videos generated by other creators using Sora. Among them, the "Split Shot" function is particularly eye-catching, allowing users to generate videos based on a series of prompt words. In addition, the "blending" function that converts photos into videos and fine-tunes text on videos also provides creators with unprecedented flexibility.
Safety and ethics are OpenAI's top priorities when launching Sora. All videos generated by Sora will have a visible watermark added by default and come with C2PA metadata indicating AI generation. Before uploading content, the platform requires users to confirm that it does not contain minors, violent or pornographic content, and copyrighted materials. Rohan Sahai, head of product at Sora, admitted that balancing creative expression and content security is an ongoing challenge.
It is worth noting that even if you do not subscribe to ChatGPT, users can still browse videos generated by others using Sora. However, in most of Europe and the UK, the official rollout of the product may take some time.
Official website address: https://sora.com/
All in all, the launch of Sora marks a significant advancement in AI video generation technology, but it also faces challenges in content security and ethics. Its functions are powerful, but the price is relatively high, and its future development is worth looking forward to.