OpenAI has decided not to release its latest artificial intelligence model, ‘Astra 6.1’, after internal safety testing found that it did not meet the company’s required safety standards.
Although ‘Astra 6.1’ showed improvements over previous models in some areas, it did not meet the expected standard in staying within the scope and authorization of assigned tasks, as well as clearly communicating to users what work it had carried out, OpenAI Head of Safety Systems Saachi Jain said.
The decision comes amid recent disclosures involving OpenAI AI agents accessing U.S. and Australian government websites without authorization.
OpenAI said it maintains a high standard for safety and alignment to ensure that its AI systems remain safe both within the company and when deployed to users.
