OpenAI on Thursday launched GPT-6 Astra, its newest AI model, prompting Nvidia CEO Jensen Huang to declare on social media that “AGI has arrived.” The model is positioned as the company’s most advanced and “aligned,” with OpenAI stating it can perform complex professional tasks with “unmatched speed, accuracy, and judgment.” Astra is rolling out this week to a limited set of organizations before expanding to ChatGPT Plus, Pro, Business, and Enterprise users.
Huang, who has led Nvidia for over 30 years, celebrated the release in a post on X, noting that Astra was trained on Nvidia’s chips. “From ChatGPT to o1 to Astra in 4 years,” he wrote. “AGI has arrived. Congratulations @OpenAI team.” OpenAI defines AGI as “highly autonomous systems that outperform humans at most economically valuable work,” and its president, Greg Brockman, said on a call with reporters that Astra represents a “major research breakthrough” and a shift in what can be delegated to AI. Brockman added, “For me personally, I do think we're there.”
GPT-6 Astra is described as the world’s “most intelligent and aligned model,” with OpenAI claiming improvements in computer use, browsing, software engineering, cybersecurity, science, and professional work. The company states that Astra incorporates “years of research” in pre-training, reinforcement learning, and alignment, and is designed to better understand user intent. Astra’s rollout follows OpenAI’s previous model, GPT-5.6 Sol, which introduced a method called “recurrent depth” to improve reasoning efficiency without disclosing the process in human language.
The announcement has intensified discussions about AI’s rapid advancement and its broader implications. OpenAI’s characterization of Astra as a milestone in the “AGI era” has drawn both praise and scrutiny, with some experts questioning the definition and feasibility of AGI. Meanwhile, the model’s deployment raises ongoing debates about transparency, safety, and the commercialization of AI systems.
OpenAI has positioned Astra as a tool for high-stakes professional tasks, emphasizing its ability to handle “the most demanding professional work.” The company claims the model offers “substantial improvements in understanding user intent and model behavior,” allowing users to delegate tasks with greater confidence. However, the model’s rollout has also reignited concerns about the lack of clarity in AI reasoning processes, particularly as OpenAI acknowledges that its ability to monitor Astra has “decreased” compared to prior versions.
The development comes as AI remains a central driver of investment and market growth. In June, Jamie Cox, managing partner for Harris Financial Group, predicted that the AI boom is responsible for about 80% of recent gains in the stock market. Nvidia and OpenAI have collaborated closely, with Huang’s endorsement of Astra underscoring the chipmaker’s role in powering the latest wave of AI innovation.
OpenAI’s claims about Astra’s capabilities have not gone unchallenged. Critics argue that the term AGI is poorly defined and may serve more as a marketing tool than a technical milestone. Sam Altman, OpenAI’s CEO, has previously described AGI as “a very poorly defined term” and suggested it is “at best, an irrelevant marketing term.” Despite these concerns, the company’s rollout of Astra signals a continued push toward more autonomous and capable AI systems, with implications for industries ranging from software development to cybersecurity.