Nvidia chief executive Jensen Huang has claimed that artificial general intelligence (AGI) has been achieved, pointing to OpenAI's newly released GPT-6 Astra model as the milestone that gets the industry there - a declaration that has already drawn pushback from AI researchers.
Jensen Huang's claim on X
Writing on X, Huang declared "AGI has arrived," citing the rapid pace of progress across OpenAI's models over the past four years. He also credited Nvidia's own hardware for the achievement, stating that Astra was trained using more than 100,000 Nvidia Grace Blackwell NVLink72 systems, with a further 400,000 Nvidia GPUs expected to come online. Huang's post also congratulated the OpenAI team on the model's release.
What can Astra reportedly do?
OpenAI introduced GPT-6 Astra on September 3, describing it as capable of handling a wide range of tasks, including computer use, software engineering, cybersecurity, scientific work and other professional applications. OpenAI President Greg Brockman has separately tied the model to the company's broader AGI ambitions, framing its launch as the start of the "AGI era" and suggesting the model could eventually be viewed as a landmark moment on the path to human-level artificial intelligence.
AGI is generally understood to mean an AI system capable of understanding, learning and performing a wide range of tasks rather than being limited to a single domain, but no single, universally accepted definition of the term exists. That ambiguity is at the centre of the current debate, with frontier AI companies increasingly describing their newest models as approaching or achieving AGI, even as many researchers question whether the underlying systems have genuinely reached general intelligence.
