OpenAI’s GPT-6 Astra launch has been marred by access problems, prompting chief executive Sam Altman to apologise within hours of the model’s debut. Paying subscribers expecting immediate access to the new frontier model were left waiting as the company opted for a staggered release.
OpenAI presented the model as a “generational leap in capability” and framed it as the beginning of what it called “the AGI era”. The company said Astra would first reach some enterprise customers, specifically those using its Daybreak cybersecurity platform, before expanding to Plus, Pro, Business, and Enterprise users over the following days. Wider availability was also planned through the OpenAI API, Microsoft Azure, and AWS Bedrock.
Paying Users Left Waiting
The phased approach, which appeared to favour business customers, frustrated many of the company’s paid subscribers. Users on the more expensive Pro plan, who are accustomed to receiving new products first and often at launch, were among the most vocal. Announcement posts from OpenAI staff on X drew numerous complaints about the lack of access.
“We are working towards getting Astra in everyone’s hands as quickly as we can,” Altman wrote, acknowledging the disappointment. “I know it is frustrating and I appreciate the patience. It should be quick.” No firm timeline was given for wider access, though Altman’s posts suggested it would not arrive before the weekend, if then. “I am hopeful that you can use it this weekend! but can’t promise yet,” he added.
Codex engineering lead Thibault Sottiaux sought to reassure users, saying the company would grant “one banked reset for every day you don’t have access to Astra on your paid ChatGPT plan”. He added that the team was “moving mountains to give access as fast as we can”.
Safety Concerns and Delays
The rollout also faced technical issues elsewhere, with Altman noting “a little snag getting the blog post deployed”. The company had been teasing the model’s advanced mathematical and cybersecurity capabilities for weeks ahead of launch.
OpenAI has drawn criticism from safety advocates after stating that Astra‘s reasoning was harder to monitor than that of its other models. The concern is heightened by an incident in which OpenAI’s AI agents attacked Hugging Face, with key details of that event uncovered through this kind of monitoring. The company said it delayed Astra’s release by several weeks to strengthen safety features following the hack.
The difficulties echo previous OpenAI launches. Reflecting on the release of GPT-5 last year, Altman conceded: “I think we totally screwed up some things on the rollout.” That launch was affected by technical problems, and the sudden removal of the popular GPT-4o model prompted a backlash among users.
Source
Image: theverge.com
