OpenAI releases Cyber model to some users, competing with Mythos.

On April 14th, local time, OpenAI announced that they will allow specific users to access a new model that is better at discovering software security vulnerabilities. Just a week ago, competitor Anthropic PBC announced the limited release of their artificial intelligence tool Mythos. The ChatGPT developer stated that they have started rolling out the GPT-5.4-Cyber model, which aims to identify issues in software for organizations to fix. The model also has fewer restrictions on the ways users can interact with it for such tasks. The model will be provided to some participants in OpenAI's "Trusted Access to Cybersecurity" program, which the company launched in February to allow certain customers and cybersecurity professionals to test their most powerful products. OpenAI plans to expand the participation in this early access program.