OpenAI is developing a new cybersecurity service for select partners to bolster digital defenses, aiming to compete with Anthropic’s recently announced Mythos model in the evolving AI security landscape.

Key Points

• OpenAI is preparing a specialized cybersecurity tool for a limited group of partners to help identify and remediate system vulnerabilities.
• The upcoming service is distinct from the company’s previously announced Spud project and does not involve the release of a new AI model.
• OpenAI currently operates a "Trusted Access for Cyber" pilot program that provides select organizations with access to advanced, permissive models for defensive research.
• The company recently projected reaching $102 billion in advertising revenue by 2030, a significant increase from its current annual projection of $2.5 billion.

Why it Matters

This development highlights the intensifying competition between frontier AI firms to position themselves as essential providers of automated cybersecurity infrastructure. As these companies race to capture market share, their ability to deliver verifiable security improvements will determine whether these tools become industry standards or remain marketing-driven initiatives.