A few days ago we had the opportunity to test Sora, the new AI video generation tool launched by OpenAI. Almost 10 months after being announced and long awaited, Sora is finally a reality.
This tool, which is part of the OpenAI ecosystem of premium services, promises to revolutionize how we create audiovisual content.
We tell you how it works, what possibilities it offers and what our impressions are after using it intensively.
What is Sora?
Sora is an artificial intelligence model that allows you to generate videos from text prompts, animate images and modify existing videos with new styles. It’s like having a creative video editor where words replace mouse clicks.
Since its initial preview in February of this year, OpenAI has worked to optimize the tool, making it faster and more accessible to a broader audience. Now, finally, it is available for users with ChatGPT Plus or Pro subscription.
Our first impressions of Sora
Accessing Sora was easy. You just need an OpenAI premium plan and register at Sora.com. In some cases you will also need a VPN, since Sora is not yet available in Europe.
Initially, we experienced some delays due to high demand at launch, but once in, the interface is intuitive and designed to inspire creativity.
The tool features a gallery where you can explore videos generated by other users, see the prompts that inspired them, and adjust them to your needs. This space is especially useful for generating ideas if you don’t know where to start.
Sora’s main characteristics
After testing several functions, these are the ones we consider the most outstanding:
Generation of videos from text
With a simple prompt like “a timelapse sunset ending with a sky full of auroras,” Sora generates a stunning video in a matter of seconds. The textures and visual details are remarkable, and the options to customize the style make each result unique.
We tried different descriptions, from natural landscapes to futuristic settings, and the tool responded with surprising fidelity to the prompts.
Storyboard
This feature is a game changer. It allows multiple AI-generated clips to be assembled into a timeline, similar to editors like Adobe Premiere Pro.
We use Storyboard to create a brief visual story, uniting several scenes inspired by an enchanted Nordic forest. Transitioning between clips was seamless, and the process was easy thanks to the user-friendly interface.
Remix and predefined styles
Another notable feature is the ability to modify existing videos with the Remix tool.
For example, we transformed an ordinary clip of a cat walking into a “film noir” style video, and then tried a “papercraft” look that gave it a stop-motion animation look.
Plus, pre-defined styles are ideal for those looking for specific results without much adjustment.
Trimming and loops
Sora also allows you to trim and loop videos directly from the platform. This is perfect if you need repeating clips for social media or creative projects.
Subscription levels
Sora functionality depends on the ChatGPT subscription plan you have:
ChatGPT Plus
For €20 per month, you can generate videos of up to five seconds in 720p resolution, with a limit of 50 videos per month. While this option is fine for experimenting, the limited duration may fall short for more ambitious projects.
ChatGPT Pro
For €200 per month, you get access to videos of up to 20 seconds in 1080p resolution, with a limit of 500 videos per month. Although unlimited “relaxed” videos can be generated that are processed during times of low server demand.
Both tiers include rendering options in three aspect ratio formats: portrait, landscape, and square, making Sora versatile for different platforms.
Sora’s Advantages and Limitations
Advantages:
- Easy to use: Even for beginners in video editing, the learning curve is minimal.
- Stunning visual results: The videos generated are of high quality and meet expectations.
- Shared creativity: The gallery and shared prompts are an inexhaustible source of inspiration.
Limitations:
- Limited duration in the Plus plan: Five seconds may not be enough for complex projects.
- Restricted access: Without a subscription, you can only browse the gallery. Additionally, it is not available in some territories. So far you cannot test Sora from Spain.
- Ethical issues: Some artists have criticized the use of their works to train models without compensation.
What’s next for Sora?
Despite his potential, Sora also raises important questions. On the one hand, there is the issue of content moderation.
OpenAI admits that there could be initial challenges in preventing misuse of the tool, such as the creation of deepfakes.
On the other hand, video generation consumes a considerable amount of computational and energy resources, which could aggravate the environmental impact of AI.
Sora: fascinating, but not revolutionary.
Sora has the potential to transform content creation for millions of people, making it more accessible and exciting. But it comes after tools like Kling, Heygen or Runway.
If you are interested in exploring new forms of visual expression or improving your audiovisual projects, Sora is an option to consider. But it is not the only one, fortunately.
But only time will tell if it becomes a standard within the world of creative AI, but without a doubt, it has already put OpenAI in the history of video generated by artificial intelligence. Maybe we could even say that the wait was worth it.
This post is also available in: