OpenAI Declares ‘AGI Era’ after Launching GPT 6

OpenAI has announced GPT-6 Astra, its new frontier model for computer use, professional work, science, coding, and cybersecurity. 

The model is rolling out initially to a limited set of organisations before becoming available to ChatGPT Plus, Pro, Business and Enterprise users, as well as through the OpenAI API and AWS.

OpenAI CEO Sam Altman said in a post on X that Astra is intended to enable “a new generation of entrepreneurship, scientific discovery, and building”, adding that the company believes it is “the best model in the world for computer use, professional work, science, coding, cybersecurity, and more”. 

Altman also said the additional time before launch was needed to meet the “safety and alignment standards required for this capability level”.

OpenAI president Greg Brockman went further during a briefing with media outlets, describing the release as the beginning of an AGI era. “Welcome to the AGI era,” Brockman said at the end of the briefing, according to Business Insider, The Information, Axios and others. 

He also said that, looking back in the future, he believed people could identify this period as when AGI was created. “For me personally, I do think we’re there,” Brockman said.

Astra has saturated several frontier evaluations, including ARC-AGI-3, ExploitBench and FrontierMath Tier 4, while setting new highs across computer-use, professional-work, coding and scientific evaluations. 

OpenAI said the model can also complete complex computer tasks faster, reporting a 1.9x improvement in task completion speed over GPT-5.6 Sol on the Mind2Web benchmark.

The model is priced at $10 per million input tokens and $50 per million output tokens through the standard API. OpenAI also offers a Fast mode at twice the standard price, with up to twice the processing speed. 

Despite the higher headline price, the company reports lower estimated API costs on several evaluations because Astra uses fewer tokens or completes tasks faster.

Astra also represents a higher level of cybersecurity capability. 

OpenAI said it meets the “Critical” threshold under its Preparedness Framework and discovered two previously unknown zero-day vulnerabilities during testing. The company has therefore added stronger safeguards and monitoring, while noting that Astra’s written reasoning is harder to monitor than GPT-5.6 Sol’s.