The declaration carries significant weight because AGI was previously tied to a contractual trigger in OpenAI’s partnership with Microsoft, though Brockman stated that is no longer the case.
By framing the release as the dawn of AGI, OpenAI is positioning itself at the definitive lead of the industry, even as critics argue the move is a marketing tactic that risks fueling public backlash.
The shift suggests that the term is being repurposed as a brand identity rather than a specific, verifiable breakthrough in machine capability.
Technically, GPT-6 Astra represents OpenAI’s largest infrastructure investment to date, utilizing over 100,000 GPUs (graphics processing units) at the company’s "Stargate" data center in Texas for pre-training.
According to OpenAI’s Aidan Clark, Astra is also the first of their models to be developed through recursive self-improvement, a process where existing AI models are used to supervise and train the new system.
This mechanism is intended to accelerate performance gains, though the company’s pivot to "AGI" branding remains controversial among those who see it as a distraction from the incremental nature of AI progress.