Sora, The AI Everyone Was Waiting For (First Look)

It happened on the first day of 12 days of OpenAI. They finally announced Sora. And not just announced, they gave us access on a brand new platform, Sora.com People from across the world except Europe and the UK can access Sora on a paid plan today. I'm extremely excited about this because you might not know this, but my background is in video production. I really started my career as a freelancer around 10 years ago just because creating videos was something I was already doing. So this release has a special place in my heart. And in today's video we'll be looking at all the details of the release. I'll review some of the features. OpenAI, Sora, let's get into it. So first things first, let me address that this is not the first video generator out there. I know a lot of you people might be tuning into the release of Sora. So let me start by contextualizing that this is one of many video generators. Actually, the Chinese video generators have been making major strides across the past few months. Here at the AI Advantage, we publish a monthly ranking where we show you what we make of these and how we rank them. And now as we're adding Sora into this, let me spoil it quickly for you, Sora. Sora is at the top as my personal number one. And most of the team agrees. And that's because it brings a blend of a video generator that is S tier on the generation quality. One might argue that maybe Runway, Gen 3 or one of the Chinese competitors are on the similar level, but I think the platform they built here, Sora.com, this is what some of the other players should have done. And frankly, only Runway got even close. It's simple, but it has all the features I always wanted to see out of AI video. We'll talk through them here in a second. But really, every week in my news roundup, I point out what new AI video features are available. But really only Runway has been shipping features consistently. This has it all. So let's level the playing ground with a few facts. Sora OpenAI state of the art video generator is available as of today across the world, with the exception of most countries in Europe and the uk. There you can use a VPN to access this, but in the OpenAI terms of service, it is outlined that your account might be banned if you do that in practice. I don't know of a single story where this happened. If you do, please leave a comment below. Alright, so let's talk pricing of this thing. First things first, free plans have no access to this. Sora is exclusively available to paid plans, which paid plans only to the plus and Pro plan. You heard this right. If you're on a teams, enterprise or education plan, you do not have access to Sora. Which is a real shame because the entire AI Advantage team is on a teams plan and nobody can use Sora there. Now I did get the Pro subscription as you might have seen during the last video. And let me just say that this is the release that made this thing thing actually worth it if you're interested in playing with Sora. So here from their blog post we have all the details of what you get with these two plans. With ChatGPT plus, for $20 a month you get up to 50 priority videos. That's the normal generation speed of the video clips. But this only includes up to 720p quality of generation, so you don't get full HD videos as you do with the Pro plan. And there's severe limitations when it comes to generating humans. This is something that's not mentioned on here, but there's Twitter discussions around it already happening. ChatGPT Pro seems to be a bit more lenient when it comes to for example, uploading your own images of humans and using the image to video features of Sora. Another thing they did not mention and I think is actually extremely significant, that the download without the watermark is only possible with the Pro plan. So yes, you do get the included generation 50 of them with your plus plan that you might already have, which is amazing. Like let's be real, they're adding anywhere from 10 to $25 of value to the plus plan to every single user. Those numbers come from having 50 generations within any of the competitors like Runway or one of the Chinese models where if you use the latest open source one and run them on rented GPUs, one generation costs $0.50. That's $25 of value added on top of what you already got with ChatGPT Plus. So that's fantastic. Let's not talk that down. But they are watermarked and this is something that did not mention. I think this is quite a big deal. If you want to use this in your YouTube video, you'll have that little Sora Waterma on the bottom. Only the Pro plan comes without the watermark, making this really pricey. In something like Runway, you can just load up $5 of credits and generate a few clips as you might need them. Here you need to pay $200 to even get a watermark free image out of this. Hey And I'm not saying it's justified, I'm just pointing out the differences here. The last interesting thing here is that with the pro plan, you get 10 times as many generations. So 50 generations versus 500. And with the pro plan, after you're done with the normal generations, what they call priority, you get an unlimited amount of slow generations. That's what relaxed video means. So the pro plan is really generate as much as you want. With Sora Plan that goes up to 10, 80, 20 seconds of duration. That's way better than most of the competitors and you can run five concurrent generations. So if you really want to play with Sora, ChatGPT Pro it is. So let's talk about the tooling, but the general story is the following. None of these tools are brand new or revolutionary. We've seen every single one of these scattered across the AI video landscape. But what they did do, in typical OpenAI fashion, is they just looked at what's available all across the market, picked out the ones that turned out to work, and implemented them into their own platform. For example, Luma Labs first came out with Dream Machine and introduced just a text to video model. Over time, they added various features with the start and end frame one being the stickiest one. Turns out it's a really great feature inside of AI video to let it transition between two objects and let the model figure out what belongs in between the starting and end point. Well, that exact feature you can find in here, they call it blend. And they even have multiple presets on how to blend something with a little bit of customization in between. But basically this was the reason most people go to Luma Labs Dream Machine or what about Runway, experimenting with a hundred different features. And one of the really sticky ones was the ability for you to loop something. You generate a clip, turn it into infinite loop. People really like that. Well, they also implemented that in here too. Again with a few variations on how you can loop something, but it doesn't end there. They have more features and yeah, they have all of the best ones. As I mentioned, check out this one. They call it Remix and basically it allows you to edit images with text prompts. So if you take this clip and you want to replace the doors with French doors, you can simply prompt that and you get this transformation very intuitively, just like a director would talk to the post production team. Or how about recut? Because in a lot of these generations you will only like a part of the video. Let's say you really love the first second, but the last Four seconds are complete trash. Well, with recut you can take that and let it regenerate the end, or you could simply use it to expand, giving you the needed control in these very very random video generators. And then the last feature I really want to point out here and that they also make a big deal out of because it is is the storyboard feature. This is actually only one company in the AI video space implemented as far as I'm aware of. LTX Studio came out with a whole suite of tools that was focused on storytelling. The caveat? The video generator was not good at all. Not even close to some of the leaders like Runway lumalabs, Dream Machine, Kling, what's of the new one Hanyuan Video or of course Minimax. The new one Hanyuan Video is actually open source and comparable in quality. Now I know some people will be out there in the comment section arguing with me that hey, Hanyuan Video when it comes to product shots in cinematic lighting has a higher quality level. Or have you considered the hyper realism in this and this type of Scene of Hanyuan vs this? It's clearly better than Sora. Maybe you can make those arguments, but I think what really matters at the end of the day is the user experience with these tools. And something like the storyboard feature, as I mentioned, has only been present in LTX Studio we now have here. Meaning you can generate one scene, edit the length, you can generate the second scene and edit all of that in a timeline where you can form a story that is coherent. This is a killer feature for most people that maybe never ever touched a video editing software. It's right in there and it's something that I pointed out many times. I wish all video generators had they got this right because they decided not to be first here, but instead to do this well and if you take all of these features that I personally already considered the best ones when it comes to other video generators, you put them into one package, you tie that up with a state of the art video generator that can generate 1080p, something that most of the other competitors cannot even do. You add folders, prompt presets, favorites, and this ability to storyboard on top. Well, I think what you get here is a package that undoubtedly creates the best consumer product when it comes to AI video. And it's not even close. And I'm confident in saying that because I've been playing with these tools on a daily basis since months now. I'm reviewing every single update and they really went above and beyond what I expected Here, they bundled up all of the best features from the competitors, paired it with their excellent model that at the very least keeps up with all of the top tier competition. And if you have a plus plan, the $20 that you were paying already arguably just doubled. And there's no need for an extra subscription with another service. Meaning that I think it's an easy prediction that this thing will steamroll the AI video market. But as per usual, we'll keep an eye on it. And later this week we'll put up a detailed comparison between all the other video generators and OpenAI Sora here so you can make up your own mind of what you make of this quality. And look, I understand that it's expensive, but let's be real, nobody that was into a video before this expected a plan with unlimited generations from OpenAI to cost any less than $200. Last week, just with 01 Pro, that was a terrible deal. Now, if you're really going to use this, you have a chance of getting the value for your money here. All right, those are my first impressions and I'll see you very soon.