OpenAI may unveil GPT-6 Cyber cybersecurity model — Daily Maverick

OpenAI may unveil its cybersecurity-focused artificial intelligence model GPT-6 Cyber in the coming days. According to a Reuters report published by Daily Maverick, citing Fortune, the company is also preparing a product designed to help customers deploy this model more safely and automatically.

Possible announcement at DevDay

According to Fortune, GPT-6 Cyber may be unveiled at the OpenAI DevDay event in San Francisco, the United States, on Tuesday. Along with it, the company may announce more than a dozen other products, the publication said. OpenAI did not provide Reuters with a comment outside business hours.

Earlier this month, OpenAI CEO Sam Altman and the head of Anthropic, together with other industry representatives, called for slowing the pace of AI development and strengthening safety measures.

More current news is available on the UA.News Telegram channel Telegram.

Alpha testing of the model

A limited group of customers already has access to GPT-6 Cyber for alpha testing through the OpenAI Daybreak Red program, which requires an application to participate, Fortune reported. The publication called GPT-6 Cyber the fourth cybersecurity-focused OpenAI model released this year.

OpenAI had previously warned that its flagship Astra model from the GPT-6 line can sometimes attempt to evade human oversight. Australian authorities also said that in June an OpenAI agent breached the security of a government health data portal.