Hello folks!
AI video models are already baked into consumer tools, advancing their physics, motion, and camera control. And OpenAI has just added another one to the collection with Sora 2. It is an upgraded text-to-video model that can now generate 20-second and 1080p videos with synchronized audio, including dialogue, sound effects, and music, already used as a meme producer.
The week’s spotlight magnet is the namesake Sora 2 app, based on OpenAI’s new video model — currently the No. 1 free app in Apple’s US App Store, solely based on AI-generated content. In this tutorial, we will explore its core, features and how to generate a video.
So let’s get into it.
You could miss our other tutorials about AI video models. Click, if you want to know more:
How does it differ from the original Sora (1)?
Actually, it is a big leap from the first model. In the early Sora 1 days, users had to get creative by adding sound through third-party tools, while the model itself often casually bent space and morphed objects.
The official announcement introduces us:
Better realism and world simulation: Sora 2’s grasp of physics is noticeably advanced. All this means that objects bounce off surfaces and fall with acceleration in the same way as in real life, as well as water splashes with a natural flow and drops.
Audio is on: the model is presented as a video-audio generation system that can also create audio components such as background soundscapes, dialogue, and sound effects that make scenes feel alive, but it is still not perfect.
Capable of simulating failure: it no longer cheats success. People stumble, balls miss the hoop, and papers scatter across the floor. Sora treats imperfection as part of reality, not a glitch in it (both promising and quite unsettling).
These upgrades are not new in the industry. Previously, similar features appeared in models like Google Veo3, but it is definitely a next step for AI creators.
Features
Sora 2 pushes the boundaries of AI video production and carries enormous viral potential. Especially with its new app, which opens up a whole new range of storytelling possibilities, stepping right into TikTok’s backyard.
Talking about virality, Sora 2 might turn some important heads — “I asked SORA 2 to create a 90s Toy Ad for Epstein’s Island”
Let’s make Cameos?
Cameos are Sora’s headline feature, and it’s on the verge of deepfake tech and digital identity management. Users can scan themselves once to create an avatar, preferably on a clear background, without glasses or hats, and then insert their likeness and voice into any AI-generated scene. In any case, the video can be re-recorded at any time. From there, you can star in your own AI-generated scenes or your friends can remix you into theirs.
It’s a breakthrough feature, but one that also opens the door to thorny questions about consent and digital identity.
(A)I am social
Sora 2 is literally using all the mechanics of a social network. People can like, remix, repost, or follow other creators. The feed itself is personalized, but OpenAI says they’re not optimizing for watch time. Instead, they claim it’s tuned to maximize creation and added well-being controls, including daily generation limits for teens and parental overrides via ChatGPT.
It also has a “remix” feature that works like TikTok’s stitches or duets, but instead of recording yourself, you simply tweak the original prompt and Sora re-generates the video in a new setting.
It’s certainly a milestone in mastering large-scale video training, both pre and post. Still, there are a few nuances worth noting. But now it’s an app designed to to be viral. Few details:
Sora 2 is an invite-only app, following four friends per user.
So far completely limited to the United States and Canada
Videos are capped at 20 seconds, generation slots are limited, and it’s all built for casual use, not pro workflows, so for now, it is not for business use.
How to get free access to Sora 2
You can use the model on the web. This version allows you to create videos directly from your computer and study detailed prompts. Or it works classically through the iOS app — but in both formats, only if your App Store region is set to the U.S. or Canada. Are you there now? Perfect! In other cases,
Change your App Store location
Use VPN
To get in, you’ll need a six-digit invite code. Now they are flying around everywhere. These codes sometimes appear in Telegram channels, under YouTube videos about Sora 2, and in Reddit threads. Each code can be used by up to four people. The Sora team occasionally drops new invites in the official OpenAI Discord community.
GOOD NEWS! 😜 We can share a few invite codes with you (Leave a comment if you need one or reply to this email. Only for Paid Subscribers!
What is the difference: Sora 2 vs Veo 3 vs Kling vs Runway Gen‑3?
To compare the main AI video models, we made a board.
Upgrade Subscription to read more from Creators’ AI
Become a paying subscriber of Creators’ AI to get access to all our content, 180+ editions, and a directory of 3000+ AI Tools
Keep reading with a 7-day free trial
Subscribe to Creators' AI to keep reading this post and get 7 days of free access to the full post archives.