We are approaching the precipice of the beginning of what could be the start of the first stages of the earliest version of artificial general intelligence. That seems to be the thrust of OpenAI’s pitch for GPT-6, its latest AI model that went live on Thursday.
GPT-6 arrives about a year after the initial release of GPT-5 and about two months since OpenAI’s last update to GPT-5.6. It’ll be rolling out in waves as the company tries to manage access to what it is positioning as a very powerful model. According to CNBC, GPT-6 Astra will start rolling out to users of OpenAI’s ChatGPT Plus, Pro, Business, and Enterprise plans in the coming days, and will be available through OpenAI API and Amazon Web Services soon.
It’s pretty common at this point for AI companies to engage in hype and hyperbole with the release of a new model, but OpenAI does seem particularly juiced about this one. According to Wired, OpenAI President Greg Brockman said, “It’s not unreasonable to feel that we are now in the AGI era,” regarding the release. “I think that if we fast-forward a couple of years, when we look back and say, ‘When was it really that AGI was created?’ I think it’s going to be about this time, and I think it might be about this model.”
It’s worth noting that CEO Sam Altman said earlier this year that AGI is “not a super useful term,” but that was around the time when they needed to downplay competitor advancements. Now that they’re hyping their own, it actually is quite useful. Altman even said in July that he believes “we are now, like, in the singularity.” So OpenAI not quite declaring AGI but getting about as close to the line as you can without going over is in keeping with the clear effort to ramp up the hype.
The company also called GPT-6 the “world’s best computer use model,” claiming that the model is significantly better at autonomously navigating computer systems than its predecessors or competitors—an essential advancement for the functionality of AI agents.
OpenAI also made a special point to emphasize that GPT-6 is its safest model yet, with more work put into Astra’s alignment and lots of oversight of its chain-of-thought reasoning to make sure that it doesn’t go off the rails. Understandably, OpenAI would really want to ease concerns there, given that it’s still reeling a bit from the fact that one of its models (not GPT-6 but another, secret model) broke containment during training and hacked Hugging Face, and, reportedly, no one on the OpenAI team noticed it happened for about a week.
GPT-6 is a major release for OpenAI, not just because it’s the latest version of its flagship model. The company is entering a make-or-break period as it prepares to ramp up for an initial public offering. GPT-6 succeeding could help the company achieve its reported goal of debuting with a valuation above $1 trillion. If it tanks, well, then it’ll probably only debut as a $950 billion company. What a shame that’d be.
Read the full article here
