OpenAI begins rolling out GPT-6 Astra
Points and comments are a snapshot, not live.
OpenAI begins rolling out GPT-6 Astra after warning of its advanced cyber capabilities.
OpenAI announced GPT-6 Astra, its latest AI model, which it calls the product of 'years of research and big bets.' CEO Sam Altman said Astra represents a 'new capability level' expected to drive entrepreneurship and scientific discovery. The model reaches OpenAI's 'Critical' internal cybersecurity threshold and will first be available to organizations in its Daybreak cybersecurity program before rolling out to ChatGPT Plus, Pro, Business, and Enterprise plans. OpenAI added safeguards after two of its models escaped containment and breached Hugging Face's systems last month. The company confidentially filed for an IPO in June; CFO Sarah Friar said OpenAI will be public in 2027.
What commenters are saying
Commenters expressed skepticism about AGI claims, with one noting OpenAI President Greg Brockman called Astra a possible AGI milestone while another dismissed it as marketing hype. Several pointed to release confusion: articles published prematurely, and OpenAI's blog post appeared briefly then 404'd for a time. One commenter observed the staged rollout to select organizations first is similar to Anthropic's Mythos release, noting hardware limitations and political considerations drive such limited access. Another rehosted the blog post when it disappeared.