OpenAI officially launched video creation artificial intelligence (AI) ‘Sora’. Starting today, ChatGPT paid users in most countries, including the US and Korea, can create a limited variety of videos depending on their plan.
OpenAI launched Sora, a text-video model, through its official blog on the ninth (local time). Currently A dedicated page called Sora.comopened the service for ‘ChatGPT Plus’ and ‘ChatGPT Pro’ users.
Sora can create videos as much as 20 seconds long using text, images, and video prompts. It takes a few minute to create a video, and a watermark and metadata indicating that it was created with AI are routinely added.
It provides editing functions resembling ▲’recut’ to chop the created video, ▲’remix’ to edit, ▲’mix’ to combine with other videos, and ▲’loop’ to repeat a certain section. . A ‘storyboard’, which redefines a selected scene as a prompt, can also be included.
Credits are required for video production. ChatGPT Plus subscribers paying $20 per thirty days are given 1,000 credits per thirty days, and Pro users paying $200 per thirty days are given 10,000 credits. Moreover, producing a 480p video costs 20 to 150 credits, 720p costs 30 to 540 credits, and 1080p costs 100 to 2000 credits. Credits don’t carry over.
Due to this fact, ChatGPT Plus can create as much as 50 videos per thirty days at 720p resolution and 5 seconds in length. Pro users can create as much as 500 videos in 1080p. The video length can also be increased to twenty seconds, and videos without watermarks can be produced.
Service on Sora.com is difficult on account of the influx of users immediately after the news was announced. The reason is, “There’s a whole lot of traffic, so it’s difficult to log in normally.”
Nevertheless, videos created by some creators or OpenAI employees are being uploaded on X (Twitter).
OpenAI asks you to comply with an agreement that states that the content you upload won’t include anyone under the age of 18, graphic or violent content, or copyrighted material. If that is violated, the account will probably be suspended.
Aditya Ramesh, vp of OpenAI, appeared on the Shipmas Day 3 stream and said, “We would like to stop Sora’s illegal activities, but we also wish to balance that with creative expression. We all know this will probably be an ongoing challenge, and it won’t be perfect the primary time. “It is probably not possible,” he said.
Meanwhile, free users may also browse AI-generated videos created by others through Sora.com. Nevertheless, it isn’t yet available within the European Union (EU). “It might take a while to roll out in Europe and the UK,” CEO Sam Altman said.
Also, today’s announcement is the third day of Open AI’s 12-day continuous event. On the primary day, the ‘o1 Pro’ and ‘ChatGPT Pro’ plans were unveiled, and on the second day, a preview of ‘o1 high-quality tuning’ was demonstrated and the official launch of next yr was announced.
In the longer term, the discharge of an update to the image creation AI Dali and the addition of a picture recognition function to GPT-4o ‘Advanced Voice Mode (AVM)’ are considered candidates.
Reporter Lim Da-jun ydj@aitimes.com
