Sam Altman apologizes for ‘messy’ GPT-6 Astra rollout that’s locked out paying users

Just hours after OpenAI launched GPT-6 Astra, CEO Sam Altman was already apologizing for what he describes as a “messy rollout” after paying users expecting access to the new frontier model were left waiting. The company hailed the model as a “generational leap in capability” on Thursday and described it as the start of “the AGI era,” a fuzzy and poorly-defined term tech executives nevertheless insist on using to promote their products. It said Astra would roll out to some enterprise customers — specifically those with access to its Daybreak cybersecurity platform — that day. This would expand to include all Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API, Microsoft Azure, and AWS Bedrock, over the next few days. The staggered release that appeared to prioritize business customers annoyed many of OpenAI’s paid subscribers, particularly those on the more expensive Pro plan accustomed to getting access to new products first, often at launch. Social media posts from OpenAI staff announcing the launch on X are filled with comments complaining about access. “We are working towards getting Astra in everyone’s hands as quickly as we can,” Altman said in an X post acknowledging the disappointment. “I know it is frustrating and I appreciate the patience. It should be quick.” Neither Altman nor other OpenAI staff online provided a timeline for when access to Astra would be expanded, though Altman’s posts suggest this will not happen before the weekend, and possibly not even then. “I am hopeful that you can use it this weekend! but can’t promise yet,” he wrote. “We will give one banked reset for every day you don’t have access to Astra on your paid ChatGPT plan, starting today,” said Codex engineering lead Thibault Sottiaux in a bid to calm concerns. “Team is moving mountains to give access as fast as we can.” There were also issues with other aspects of Astra’s rollout, which the company has been teasing with updates on its advanced mathematical and cybersecurity capabilities for weeks. Altman said there was “a little snag getting the blog post deployed.” OpenAI has also faced intense criticism from safety advocates after it said Astra’s reasoning was harder to monitor than its other models. This appears to be a growing trend at the frontier of AI development but considered especially concerning in light of OpenAI’s AI agents attacking Hugging Face, about which crucial details were learnt through this kind of monitoring. OpenAI said it delayed Astra’s release by a number of weeks to bolster safety features in the wake of the hack. Other big OpenAI releases have also been marred by missteps and frustration. “I think we totally screwed up some things on the rollout,” Altman said last year when reflecting on the company’s last major release, GPT-5. The launch was plagued with technical issues and the decision to suddenly remove the popular GPT-4o model sparked outrage among users.