OpenAI has released Astra, which President Greg Brockman describes as the world's most intelligent and aligned model and the start of the artificial general intelligence (AGI) era. This announcement comes as the company also releases GPT-6, positioning it as a key milestone in a decade-long push toward AGI. However, CEO Sam Altman has previously described AGI as an irrelevant marketing term.
The company claims Astra represents a new frontier in computer and browser use, offering unmatched speed, accuracy, and safety. The model excels at coding and computer use, and it is capable of spotting more security vulnerabilities than the most advanced OpenAI model publicly available today.
The release of Astra follows a pause in its training caused by a serious AI safety incident involving other models under development that sparked international concern. While outside experts worry about whether the company has effective safety and security controls in place, OpenAI stated it has built in safeguards to mitigate security risks. Access to Astra will be provided first to companies participating in an application-based cybersecurity program.