OpenAI’s GPT-6 Astra Debut Stumbles Before Reaching Paying Users

OpenAI’s GPT-6 Astra Debut Stumbles Before Reaching Paying Users
OpenAI launched GPT-6 Astra on Thursday with sweeping claims: Sam Altman called it the company’s best model for “computer use, professional work, science, coding, cybersecurity, and more,” promising a new wave of entrepreneurship and discovery.

The rollout, however, began with a narrower audience than the rhetoric suggested. Astra initially went to selected enterprise customers with access to OpenAI’s Daybreak cybersecurity platform, with broader access promised in the following days. OpenAI had delayed the release by weeks to strengthen safety features, amid concerns that the model’s reasoning was harder to monitor, according to reporting on the launch.

That staged release quickly angered paid ChatGPT subscribers—especially Pro users accustomed to early access. Within hours, Altman conceded the mismatch between expectations and delivery: “first, sorry for the messy rollout,” he wrote, saying broad access for API customers and subscribers should begin “in the near future.”

OpenAI’s response was both apologetic and practical. The company said it would issue banked resets for each day paid users lacked Astra access, while Altman said the team was working to make the model available quickly. The episode echoed previous major launches that drew complaints over technical trouble and abrupt product changes.

Later, the company’s timeline finally caught up with its announcement. Altman said Astra had reached Pro, Enterprise and Business Premium users through Work/Codex and the API, before the rollout moved to Plus and Business customers. He then declared: “Now out to all Plus and Business users. Happy building!”

For OpenAI, Astra’s debut ended in expanded availability. For subscribers, it was a reminder that frontier-AI launches can still arrive in tiers—even after the applause starts.

Continue reading https://foxvector.com/stories/01a06fa5-02c2-06d2-73b1-11aa0221b751