On September 7, 2026, OpenAI officially released the large-scale model Astra, claiming that it is the most intelligent model in the world and aligned with human values. Nvidia CEO Jensen Huang congratulated and issued a statement stating that the era of AGI (Artificial General Intelligence) had arrived. OpenAI President Greg Brockman welcomed this new era during a subsequent teleconference. However, there are disputes over this view: Gary Marcus, from the AI Research Institute and a critical scholar, argued that Huang’s judgment lacks scientific definition and evidence, and that merely announcing the takeover of the scientific realm is insufficient. He pointed out that Astra currently meets only one or two of the 10 AGI standards proposed. Additionally, OpenAI CEO Sam Altman said in a podcast that AGI is a vague term and may even be just a “unimportant marketing term.”
Nvidia CEO Jensen Huang congratulated OpenAI on the release of the large-scale model Astra, stating that “AGI has arrived”. On Thursday, OpenAI officially released Astra, calling it the world’s “most intelligent and human-values-aligned model” capable of completing the most challenging professional tasks with unparalleled speed, accuracy, and judgment. Greg Brockman, President of OpenAI, said during a teleconference: “Welcome to the era of AGI.” In response, AI research institute and critic Gary Marcus argued that Huang’s announcement was premature, pointing out that he had not provided a scientific definition or evidence, and merely declared unilaterally that AGI would take over scientific issues. Marcus also presented his own 10 criteria for defining AGI, indicating that Astra currently met only one or two of them. Additionally, Sam Altman, CEO of OpenAI, stated in a podcast that AGI is a term with vague definitions and may even be just a “meaningless marketing term”.