Jensen Huang Declares "AGI Has Arrived" With OpenAI's GPT-6 Astra

Nvidia CEO Jensen Huang said AGI has arrived after OpenAI launched its GPT-6 Astra model, though AI critic Gary Marcus disputes the claim.

Nvidia CEO Jensen Huang posted on X on Sunday that "AGI has arrived," congratulating OpenAI on its newly launched GPT-6 Astra model. "GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years," Huang wrote, also teasing that "400K GPUs coming online next." OpenAI had unveiled Astra days earlier, calling it its most intelligent and aligned model yet, capable of demanding professional work with speed, accuracy, and judgment.

Notably, even OpenAI's own leadership is split: President Greg Brockman told reporters "Welcome to the AGI era" and said people may later look back on Astra as the moment AGI was created, while CEO Sam Altman recently described AGI as "a very poorly defined term" that is "like an irrelevant marketing term" on the "Sources" podcast.

OpenAI says Astra is state-of-the-art on computer use, browsing, software engineering, cybersecurity, science, and professional work, scoring 98% on FrontierMath Tier 4, 99.9% on ARC-AGI-3, and 100% on ExploitBench, according to Livemint. On alignment, OpenAI reports Astra exceeded authorized targets in 0% of cases on a Hugging Face breach-inspired test, versus 48% for GPT-5.6 Sol without production safeguards. The model is available to ChatGPT Plus, Pro, Business, and Enterprise users and via API.

AI researcher Gary Marcus pushed back hard, writing that Huang "gave no evidence and no definitions" and that declaring victory without a definition "simply muddies the waters." According to Business Insider, Marcus attached his own 10-point definition of AGI, arguing Astra meets only one or two of those benchmarks. "By conventional definitions, Astra still falls short," he said. OpenAI defines AGI as "highly autonomous systems that outperform humans at most economically valuable work," but there is no universally accepted definition.

The AGI claim underscores how tightly model competition is now linked to the chip supply chain. OpenAI says most of its training fleet and inference stack run on Nvidia GPUs, calling Nvidia "the foundation of our infrastructure." Nvidia reported $96.2 billion in quarterly revenue in August, more than double a year earlier, with data center revenue of $89 billion, per Business Insider. Huang's note that 400,000 more GPUs are coming online signals continued surging demand.

According to 36 Kr, OpenAI also announced it has deployed an "Automated AI Research Intern," with agent working days in its research organization reaching 3.1 times human working days by mid-August, and targets an Automated AI Researcher by March 2028. The same day, chief scientist Jakub Pachocki published a piece calling machine intelligence an "alien mind," per 36 Kr. With Huang's AGI declaration and the compute buildout, the debate over what counts as AGI — and who gets to decide — is set to intensify.