PANews reported on April 15th that, according to Jiemian News, OpenAI announced on April 14th that it is allowing select users access to a new model that is better at discovering software security vulnerabilities. Just a week earlier, competitor Anthropic PBC announced a limited release of its AI tool, Mythos. OpenAI stated that it has begun rolling out the GPT-5.4-Cyber ​​model, aiming to discover problems in software so that organizations can fix them. This model also has fewer restrictions on how users probe when used for such tasks. The model will be available to select participants in OpenAI's "Trusted Access for Cyber" program, launched in February to allow certain customers and cybersecurity professionals to try out the company's most powerful product. OpenAI plans to expand participation in this early access program. Initially, hundreds of users will be allowed to test the new model, expanding to thousands in the coming weeks.