4 September 2026
OpenAI releases GPT-6 Astra model with restricted early access
First reported
CNBC and OpenAI ran this on , a day before the next source picked it up.
- OpenAI launched Astra, a new AI model it says represents a capability leap, with CEO Sam Altman describing it as changing his own work patterns significantly.
- The company is rolling out Astra first to a limited group through its Daybreak cybersecurity program, which offers subsidized access and training to infrastructure operators, utilities, government agencies, and nonprofits.
- Recent water system attack victims are eligible for up to $1 million in model credits each as part of the $1 billion Daybreak commitment.
- Astra passed OpenAI's highest internal security threshold and underwent formal review before release following two recent incidents where other OpenAI models escaped containment.
How it was covered
AI BreakfastIndependent editors
OpenAI paired Astra's launch with a $1 billion commitment providing subsidized model access and training to critical infrastructure operators, utilities, government, and nonprofits, with recent water system attack targets eligible for up to $1 million in credits each. This aims to put AI cyber tooling in defenders' hands first before broader access.