[OpenAI Launches Cyber to Compete with Anthropic, AI Vulnerability Exploitation Raises Regulatory Concerns] BlockBeats News, April 15, according to foreign media reports, OpenAI is allowing a select group of users to access a brand-new AI model, GPT-5.4-Cyber. The model aims to identify issues in software so organizations can address them. A week ago, competitor Anthropic announced the release of an AI tool called Mythos, which is reportedly designed specifically to identify and exploit vulnerabilities in operating systems and web browsers. Mythos has quickly raised concerns among financial institutions and government agencies. At a meeting convened last week by Besant and Powell with Wall Street leaders, executives were warned to take Mythos seriously. OpenAI plans to increase the number of participants in this early access program. It will allow hundreds of users to test the new model, then expand to thousands in the coming weeks. (Jin10)