Sora, the AI Video tool that was teased so long ago is finally available! You can use Sora at https://sora.com
Here’s a tidied version of the transcript without timestamps:
The day is finally here! We’ve been hearing about this for a long time, but OpenAI has finally released Sora, their AI video generator model. I’m super excited to play with it. I haven’t personally used it yet, but it took so long to come out that all the other AI video generators have improved and almost reached Sora’s level. However, from what I saw in today’s demos, Sora still seems to have an edge over many other tools.
Most people are probably asking how much Sora costs to use. If you have a paid ChatGPT account, you can use it with your account. The $20 a month Plus account allows you to generate up to 50 videos per month (1,000 credits) at 720p. If you upgrade to the $200 a month Pro Plan, you can generate up to 500 videos per month (10,000 credits), get unlimited relaxed videos, generate at 1080p resolution, and have five concurrent generations. You can also download videos without a watermark on the Pro Plan.
It’s important to note that Sora is not available in every country yet; there are some legalities being sorted out in regions like the UK and the European Union due to stricter AI-related policies.
When I logged into the dashboard, I found a familiar and intuitive interface. The left menu includes options for recent videos, featured content, saved items, and more. There are filtering options by prompts and aspect ratios, as well as settings for username and email.
I started testing video generation with prompts like “a monkey on roller skates.” This has been a challenge for other video generators so far. After generating the first video at 480p for speed, I quickly added more prompts like “a wolf howling at the moon” and “a cat eating tacos.” Unfortunately, one of my prompts failed due to an unexpected error.
The first successful generation was the monkey on roller skates. Although it was only 480p, it was the best rendition I’ve seen so far. The quality showed some pixilation due to the low resolution, but overall it was impressive.
As I continued testing with various prompts and presets like “balloon world,” “stop motion,” “archival,” and “film noir,” I found that while some generated decent animations, others struggled with specific actions like juggling or morphing.
One feature I found particularly powerful was the storyboard option, which allows multiple prompts along a timeline. However, even with detailed prompts like “a man standing on a cliff morphing into a wolf,” the results were not as expected.
Overall, Sora shows great potential but currently struggles with certain animations and has long wait times due to high traffic. Many users are experiencing delays of up to an hour for video generation.
As more users create videos with Sora, impressive examples are emerging in the featured section. The quality of these videos indicates significant advancements in AI video generation compared to earlier models.
In conclusion, while there are some issues with wait times and prompt effectiveness, Sora represents an exciting step forward in AI video technology. I’m looking forward to exploring its capabilities further and sharing my experiences on social media as I continue testing it out.
