I Paid For OpenAI’s Sora So You Don’t Have To!
What is Sora?
Sora is OpenAI’s recently released text to video AI generator. When Sora was first unveiled to the world in 2024, most video professionals were unsettled. We knew AI was coming but the fact that it was already this good was shocking.
December 2024 they opened it to the public at a price of $20 for use with limited credits, or $200 with the ability to generate unlimited 1080p videos up to 20 seconds.
I spent a few hours playing around with it (see the full demo at the bottom of this post) and have to admit I’m intrigued.
How Does it Work?
Sora allows you to use text-to-video prompts to create almost anything you can imagine.
Just like text-to-image prompt tools, it has the same troubles. People look uncanny, physics are usually completely broken, and objects appear at random.
The best practice is to take a look at what’s trending to get an idea of the prompts to use. You don’t want to be too specific in the wrong things, but you also need to paint a picture in a specific way to get a useable shot.
How Much Does Sora Cost?
It takes time to get the hang of, and the credits go by quickly.
As of January 2025, $20 buys you 1,000 OpenAI video credits, roughly 50x 480p 5-second generations per month. Each is about 25 credits. This tier is also limited to 720p creations and 5-seconds.
The higher tier unlocks HD video, and also allows you to remove the watermark. You are allowed unlimited generations, with 10,000 credits at the priority speed (it is VERY fast). They claim you also can have unlimited video generation with deprioritized render times.
I have to imagine these prices will go up if users are maxing out the higher tier – which presents a great value but is also expensive if you are a power user.
How to Use Sora
I’ve taken the time to put the Sora through some paces with the consideration of how we might use it as a creative tool.
It’s nowhere near replacing what we do. But as way to create some quick visuals for content, or to pitch or experiment with ideas, it is absurdly good.
I took the time to make a little demo if you want to see how it works.
Is Sora Worth Using?
Yes. If you can stomach the idea.
The model delivers amazing images when it’s working right. It’s only going to get better.
For creators – where the ability to produce and scale ideas on cheap or even free provides a lot of value – this will be highly used.
As video producers or filmmakers we are concerned about how our work is valued and used. When it comes to the ethical implications of using AI at this scale it’s a different story.
I personally am deeply ambivalent. We don’t know how this model has been trained, but it’s safe to say our own work on the internet has been used. The advertising and film industry in 2024 felt massively disrupted, and I can’t help but feel a lot of work was flowing towards AI.
And yet I can already see so many applications to enhance our workflow and creative development. So most likely we will have to use it to stay competitive.
The Insane Amount of AI Video We’re About to See
We’re living in an LLM tidal wave already. The outcome from this on the internet and ways we work are a big unknown.
If the future is one big homogenized average of words, ideas, and thought–count me out. As of now it’s wait and see.