OpenAI Launches Cyber Focused on Anthropic, AI Vulnerability Discovery Capability Raises Regulatory Concerns

BlockBeats News, April 15th, according to foreign media reports, OpenAI is allowing a selected group of users to access a brand-new artificial intelligence model, GPT-5.4-Cyber, aimed at discovering issues in software for organizations to address. A week ago, competitor Anthropic announced the release of an AI tool named Mythos, reportedly designed to identify and exploit vulnerabilities in operating systems and web browsers. Mythos quickly raised concerns among financial institutions and government agencies.

During last week's meeting with Wall Street leaders convened by Bernett and Powell, executives were warned to take Mythos seriously. OpenAI plans to expand the participants in this early access program. It will allow hundreds of users to test the new model, then scale up to thousands in the coming weeks. (FXStreet)