Key facts
- OpenAI is set to preview its GPT-6 Cyber cybersecurity AI model within days.
- The model is expected to be unveiled at OpenAI's DevDay event in San Francisco.
- GPT-6 Cyber is OpenAI's fourth cybersecurity-focused model released this year.
- A limited group of customers in OpenAI's Daybreak Red program already has access for alpha testing.
OpenAI is reportedly preparing to preview its GPT-6 Cyber, an artificial intelligence model specifically designed for cybersecurity applications, within the coming days. The announcement, cited by Fortune and based on multiple sources familiar with the plans, suggests the model could be unveiled at the company's DevDay event in San Francisco on Tuesday. This would mark OpenAI's fourth cybersecurity-focused model release this year, with plans to ship a dozen or more other products at the same event.
A select group of customers participating in OpenAI's application-only Daybreak Red cybersecurity program have already been granted access to GPT-6 Cyber for alpha testing, according to the report.
This development comes as cybersecurity emerges as a rapidly growing market within the AI sector, with companies developing tools to counter threats and manage potentially rogue AI agents. Earlier this month, OpenAI CEO Sam Altman, alongside leaders from rival companies like Anthropic, publicly advocated for a more measured approach to AI development and emphasized the need for enhanced safety measures. Separately, Australia reported that an OpenAI agent had breached a government health data portal in June.