The video creation model Sora, developed by OpenAI, an artificial intelligence firm financed by Microsoft, has been completely rolled out. It can produce a new video with input in the form of text, images, and video. However, it is not incorporated into ChatGPT; it is a stand-alone service. The video creation tool functions independently of ChatGPT and is available via a different website. Nevertheless, a ChatGPT Plus or the recently released Pro membership is needed to access Sora.

Additionally, OpenAI unveiled Sora Turbo, a faster variant that, according to the firm, performs noticeably better than the initial Sora model that will be debuted in February 2025.

What’s new with Sora from OpenAI?

According to OpenAI, premium members can create videos with 1080p quality, a maximum duration of 20 seconds, and widescreen, vertical, or square aspect ratios. Sora allows users to extend, remix, and blend existing or create whole new works from words by accepting text, photos, and videos as input suggestions.

A storyboard tool in the revised Sora interface lets users precisely define inputs for every frame. Along with “All Videos,” “Favourites,” and “Upload” sections for simpler navigation, the interface also includes sections like “Featured” and “Recent” feeds that highlight user-generated content.

Subscriptions and availability

ChatGPT Plus and the recently launched ChatGPT Pro memberships include access to the Sora video generator. Users can create up to 50 videos using ChatGPT Plus, each lasting up to five seconds and with 720p resolution.

Additional features available with the Pro subscription include five continuous generations, the ability to create up to 500 movies, and 20-second videos in 1080p definition. Pro members can also download videos without watermarks.

Limitations of Sora from OpenAI

Safeguards have been put in place by OpenAI in the same way as safety that is formulated by Google’s video generative AI models, such as the default visible watermarks on videos and an internal search tool that uses technical attributes to confirm the origins of content.

OpenAI is now limiting submissions of human-facing content to prevent abuse, especially with materials like deepfakes. However, as deepfake mitigation techniques are improved, the company intends to make this capability available in the future.

Author: mayank

Mayank Firani is a dedicated blog writer with expertise in crafting compelling content across various niches.