ChatGPT Manufacturer OpenAI introduced Sora, its AI engine for converting text prompts into videos, back in February 2024, but we had to wait until December 2024 for the full release, and even then it was only available in the US. So far the results we got from this were amazing.
Sora works exactly as you’d expect. You simply enter what video you’d like to watch and Sora does the rest. There are some restrictions on the length of video clips and their number. A ChatGPT Pro subscription gives you more credits and allows you to create longer clips without watermarks.
Here we’ll explain everything you need to know about OpenAI Sora: what it can do, how it works, and how to get it. The era of AI-powered text-based filmmaking has finally arrived.
Release date and price of OpenAI Sora
In December 2024, Sora received its first public release as part of OpenAI”12 days OpenAI‘ event. Sora requires a subscription to ChatGPT Plus, Teams, or Pro to use and is currently only available in the US, but we expect it to be released worldwide at some point in 2025.
WITH ChatGPT Plus ($20/£16/AU$32 per month) or Teams ($25 US/£20/AU$40 per month per user) you get all the benefits ChatGPT Plus user plus up to 50 priority Sora videos (equivalent to 1000 credits) per month in 480p resolution. Your videos can be 480p or 720p and 5 seconds long. They also have a watermark.
WITH Subscribe to ChatGPT Pro ($200/£165/AU$325 per month) gets you unlimited access to ChatGPT 4o and o1, plus limited access to the latest version of o1 pro mode. Plus, you get up to 500 priority videos (equivalent to 10,000 credits). You also get unlimited “casual videos” that will take longer to create than priority videos. Videos are up to 1080p resolution, 20 seconds long, and you can create five videos at once. You also have the option to download videos without watermarks.
Given these limitations, it’s pretty clear that if you want to use Sora for any professional purposes, you’ll need the more expensive Pro subscription.
What is OpenAI Sora?
You may be familiar with generative AI models such as Google Gemini for text and Dall-E for images, which can generate new content based on huge amounts of training data. For example, if you ask ChatGPT to write you a poem, what you get back will be based on the many poems that the AI has already learned and analyzed.
OpenAI Sora is a similar idea, but for video clips. You give it a text prompt, such as “a woman walking down a city street at night” or “a car driving through the forest,” and receive a video in response. As with AI image models, you can be very specific about what should be included in the clip and the style of footage you want to see.
To better understand how this works, watch some example videos posted by OpenAI CEO Sam Altman – shortly after Sora was introduced to the world, Altman replied to clues posted on social media, bringing back videos based on text like “a wizard in a pointed hat and blue robe with white stars casts a spell that shoots lightning from his hand while holding an old tome in his other hand.”
How does OpenAI Sora work?
At a simplified level, Sora’s technology is similar to that which allows you to search the Internet for images of a dog or cat. Show the AI enough photos of a dog or cat, and it can detect the same patterns in new images; Likewise, if you train an AI on a million videos of a sunset or waterfall, it can generate its own.
Of course, there is a lot of complexity behind this, and OpenAI has provided deep dive about how his artificial intelligence model works. It’s trained on “internet-scale data” to know what realistic videos look like, first analyzing clips to know what it’s looking at, and then learning to create its own versions when asked.
So, ask Sora to make a fishbowl video and he will come back with an approximate result based on all the fishbowl videos he has watched. It uses what are called visual patches – smaller building blocks that help the AI understand what should happen where and how different elements of the video should interact and develop, frame by frame.
Sora is based on a diffusion model, in which the AI starts with a “noisy” response and then works on the “clean” result through a series of feedback loops and predictive calculations. And like other generative AI models, Sora uses transformer technology (the last T in ChatGPT stands for “Transformer”). Transformers use a variety of sophisticated data analysis techniques to crunch through reams of data—they can understand the most important and least important parts of what is being analyzed, as well as figure out the surrounding context and relationships between those pieces of data.
What we don’t fully know is where OpenAI found its training data—it didn’t say what video libraries were used to power Sora, although we know it has partnerships with content databases. for example Shutterstock. In some cases you may see the similarities between the training data and the output that Sora produces.
What can you do with OpenAI Sora?
When you open Sora, you see a landing page that is a grid of video content created by all the other users. This is a great place to start if you’re looking for inspiration, because you can use any of these videos as a starting point for your own creations using what OpenAI calls a “remix.”
When remixing, you can make a light remix, a subtle remix, a strong remix, or even your own remix. Alternatively, you can create your own original video based on, for example, the prompt “Man Dances on the Moon with a Stetson Hat.”
Sora also has a storyboard tool to help you create and direct the type of video you want. You can use it to precisely specify the input data for each frame. There are also presets to help you create videos, such as Film Noir and Stop Motion.
OpenAI currently acknowledges some of Sora’s limitations. Physics doesn’t always make sense: people disappear, transform, or merge with other objects. Sora doesn’t map out the scene with individual actors and props, she makes an incredible amount of calculations about where the pixels should go from frame to frame.
In Sora’s videos, people may move in ways that defy the laws of physics, or details (like biting into a cookie) may not be remembered from one frame to the next. OpenAI is aware of these issues and is working to fix them. Some examples can be seen on the page OpenAI website Sora to understand what we mean.
Despite these mistakes, going forward OpenAI hopes that Sora can evolve into a realistic simulator of the physical and digital worlds. In the coming years, Sora’s technology could be used to create imaginary virtual worlds for us to explore, or allow us to fully explore real-world locations that are rendered using AI.
How to access OpenAI Sora?
At this time, you cannot sign in to Sora without a Plus, Teams, or Pro subscription. You access it from sora.com. Sora is also only available in the US at the moment, so if you try to access it from another country, you’ll see a message saying it’s unavailable.
OpenAI has a history of eventually getting all of its paid products into the free tier somehow, and we expect a version of Sora to eventually be released in the free ChatGPT tier, but it looks like we’re a long way off from that yet. it’s happening right now.
Is there anything else like Sora?
Since Sora was announced, several alternatives have been released. runway is one of the most popular video tips generators. He has Runway Gen 3 Alpha there is a version available that you have to pay to access ($144 per year, or about £111 or AU$214), but it Gen 2 runway version available for free trial. You get 525 credits per month, but are limited to 4-second clips.
Google has its own video invitation generator called I understandwhich is expected to launch later this year. It certainly looks impressive, but you’ll need to get on the waitlist to try it out.
Another contender for the throne of the AI video generation is Pike. It’s designed for individuals, and you can upload and customize characters, objects, and settings using the new Scene Ingredients feature.
Finally, Luma AI Dream car is another contender for the throne of text-to-video conversion. Again, you simply enter text prompts and the program generates a video. You can subscribe to free account but due to high demand, the free tier is limited to 20 generations per day. Luma also offers paid tiers costing up to $399.99 (about £309 and AU$595) per month for 2,000 generations and the highest queue priority.