SORA-2 Video Tips

Plus invite codes for you 🥳

October 06, 2025

SORA-2 came out last week and it’s the hottest thing since VEO-3.

Admittedly, SORA-2 is brand new, so very few people have done full-on deep dives into how to produce the absolute highest quality videos.

But for now, we’ve dug up quite a few tips from multiple sources.

So I thought I’d organize them for you and send the list over.

*If you’re still struggling to get an access code, jump into our Power Users community and we’ll make sure we get you one.

Anyways, here are the tips:

#1 - High Quality Onboarding Video

Unlike VEO-3, with SORA-2 you can’t upload a photo or video of someone to use as the jumping-off point for your video.

Instead, SORA uses what are called Cameos (meaning you and one or more other people all appear in the same video).

The way Cameos work is you record a video of yourself (including you speaking) during the registration process so that Sora can make videos that look and speak like you.

While this might sound obvious, it’s critical you take this video somewhere that:

Has good lighting
Is positioned against a contrasting background (ideally green screen)
Is dead silent (so the AI picks up nothing but your voice)

As a side note, the Cameo feature can be turned on and off.

So if you don’t want other people doing cameos using your face/voice, don’t worry.

With that said, you can only do cameos with other people if you leave it turned On (which is how people have been doing these Sam Altman videos lately).

Either way, we think this is a super fun feature. And if you and your friends create solid onboarding videos, you can produce some very high-resolution duos together.

#2 - Cartoon Videos Work Well

Whether it’s Minecraft, anime, South Park, or something else, Sora videos that leverage highly established visual styles are generating highly accurate outputs.

I mention this because regular old human videos can be hit or miss. Some are surprisingly high resolution. Others are flat out terrible.

Admittedly, you might not have any use cases for creating an Avatar or Sponge Bob themed video. And that’s fine.

But if you do, your likelihood of getting a high-quality video is higher.

#3 - Understand Its Limitations

Unlike VEO-3, Kling, etc., Sora has a very small prompt window (2,500 characters max).

The reason for this ties in with Sora being somewhat of a social app.

With VEO-3, which is connected to Google’s Flow software, it’s possible to produce cinematic-style videos and storyboard them with a full-blown video editor.

Sora, however, is not that kind of tool.

Instead, OpenAI wanted people to be able to get value out of it without having to write super elaborate prompts. Specifically, they wanted it to produce high-quality videos based on Natural Language Prompts.

So, if you’re an AI Power User, odds are you’ll need to tone your prompts down.

Second, Sora videos are a max of 10 seconds long with a max resolution of 720p.

Under the Pro plan, however, that gets bumped to 20s and 1080p.

#4 - Sora Specific Prompting

According to OpenAI, Sora prompts should be ~120 words (2,500 character max) and contain the following elements:

Focus on one scene and one core action
Outline a clear “Subject + Action + Setting” combo
Contain simple camera/style cues (e.g. cinematic close-up)
Optional: Include sound cues if desired

Example below:

The bad news is that, unlike VEO-3, SORA won’t let you change scenes mid-video.*

*Of course, you can still create multiple individual videos and string them together.

The good news is that, relative to the $2.50 to $5 it cost to create ONE video with VEO-3 when it first came out, Sora gives you 50 every 24 hours (for FREE).

*If you want access to Sora but have not been able to get an invite code, join our Power Users community and we’ll get you inside.

In conclusion, Sora is an interesting evolution in the world of AI video.

By making Sora a social app, OpenAI makes it clear they’re targeting the masses (versus something like Kling or Higgsfield, which appeal to power users who desire cinematic quality video).

Admittedly, there are some rumors OpenAI is taking the Venture Capital approach, where they get as many people hooked on the app as possible before starting to charge for something that used to be free/low cost (similar to what Uber, Doordash, etc. did).

And there are rumors OpenAI is being crushed under the weight of how much processing power all this free video generating requires. So it remains to be seen what their long-term plan is.

For now, I’m already seeing some Sora viral videos on Instagram (I don’t use TikTok).

So our opinion is: Might as well enjoy it while it lasts!

*Interested in learning more about the most cutting edge AI tactics?

Make sure to follow me, Sentient’s Head of Product on X…

And join Sentient’s brand-new AI Power Users community (it’s free : )