NEW GPT 5.6 is INSANE!

**Julian Goldie** (0:00)
So, GPT 5.6 is here, but not for everyone. And that's what I'm going to talk about today with the new release of GPT 5.6, plus what's going to happen of whether we're entering a new era where essentially you have to be approved for AI models before you can actually get access to it. So, right now, the US government apparently has actually asked OpenAI to stagger the release of GPT 5.6. And this is obviously after the whole Fable 5 incident where Fable 5 literally got shut down and taken down after like three days of being released. It's powerful it was. And so, what we're seeing here is an approval by customer situation. And here's exactly what's happening right now. So, essentially, you can see the article here from Axios, right? And it says the administration, Trump administration asks OpenAI to limit next model release, which is pretty wild. So, essentially, what's going on here is there's a staggered rollout of GPT 5.6. Now, whether it will get rolled out to everyone, eventually, who knows? They're also looking at a 2027 IPO, as you can see in this article by Silicon Angle. And essentially, the rollout is going to be for people and just a small group of partners rather than for the whole public. And so, what's happening basically is like, Washington are now reviewing frontier artificial level intelligence models before they actually ship widely.
Pretty wild stuff.
So, this was disclosed by CEO Sam Altman during a staff question and answer session on Wednesday as detailed in a report by The Information, which you can see right here. And so, essentially, this means the people have to be approved. It's getting rolled out to a small group of people. And due to the Fable 5 situation, it seems like we're rolling into an era where frontier models aren't available for everyone. They're basically going to go out to small partners at first. And then maybe once they've been fully tested, then potentially everyone gets access. Another rumor is that maybe in the future with frontier models, you have to be approved to use them before you can actually get access to them. But either way, this is pretty well. Bear in mind as well, if you look at, for example, China, you got Glm 5.2 and KimiK 2.7. They both came out and they're both pretty powerful. I mean, especially Glm 5.2.
It's a setup where essentially you've got a frontier level model, but the weights are open. It's an open source project.
Even if you have a super powerful setup, you can actually run it locally for free, which is pretty crazy in itself.
If you look at this situation, it's like, okay, well, Chinese models are open source and they're getting at the point where they're at the same level or even potentially in the future better than stuff like Claude Opus 4.8. And then in the US, you're looking at frontier models that are being blocked or being staggered in terms of their rollout or just not available to everyone in the future. So it's a weird sort of situation we're getting into. You can see in this tweet by Stephanie Palazzolo, for AI companies on the verge of releasing cutting edge new AI models, there's a new normal in the wake of Trump's administrations, showdown with Anthropic. So essentially during a Q&A session with OpenAI and CEO Sam Altman, the company would only release its model in a limited preview to a small group of partners. And this has never really happened before, to be honest. Like there's been some security models that come out, and you've seen that with Mirfos as well. But this is really, really interesting. There's never been a situation like this. And there's two things that have happened recently. So Fable 5 obviously got shut down after three days, which has never happened before. And then these models have been rolled out to just small groups of testers, probably security companies, I would imagine.
So essentially during an OpenAI memo reported by Silicon Angle on the 25th of June, the government would be approving access customer by customer during this preview period. Now, there is a wider rollout that's hoped for a couple of weeks later. So we could be looking at like mid-July when everyone gets access to July to GPT 5.6. But who knows what's going to happen next. Also, Altman apparently he made it clear that he isn't happy about it long term.
Apparently in the memo, this is what's being reported. This has not been confirmed by OpenAI publicly. He said the company has said this is not our preferred long-term model and that they'd worked toward a more sustainable approach for future releases. So if you look at this, how a gated launch actually works is it gets built, then it gets vetted, then it gets approved one customer at a time, and then there would be a wider release a couple of weeks later. And this is really the first time a US model launch has a queue. So we're used to AI launches being instant and global, a new model drops and within the hours everyone's testing it. That's a rhythm the whole industry has previously run on. But GPT 5.6 would break that rhythm. So for the first time, a major US model is shipping behind a managed line. And also the bigger worry isn't really about this one single launch, it's really that it's a precedent, right? Because if new models are signed off on a customer by customer basis, that might become the template in the future. So for example, you might be looking at the next models like that, or the next one after that, or from every single lab, especially in the US.