OpenAI's secret roadmap to superintelligence has just been leaked, and they've introduced a new system called the AI preparedness framework to monitor and assess progress towards human level AI.

What does it do?

So this framework aims to evaluate AI systems across 12 key capabilities on a five-level scale, ranging from chatbots to systems capable of running entire organizations.

And it gets more interesting, OpenAI believes it is already close to solving level two with AI that can reason almost as well as someone with a PhD.

What we do know is that new leaks have revealed the existence of strawberry models, which will allow the AI to not just generate answers to queries but to also plan ahead enough to navigate the internet autonomously and reliably to perform deep research.

But let's be honest, with the Sora and the Omni models still unreleased, we can't expect public access to this advanced model anytime soon.