Openai: OpenAI has announced the availability of a new artificial intelligence (AI) model dedicated to cybersecurity, a move that comes just days after rival Anthropic unveiled a similar model named Mythos. The developer of the ChatGPT application said it has begun rolling out the GPT-5.4-Cyber model to identify software vulnerabilities, enabling companies and organisations to address them early. The company explained that the new model offers users greater flexibility in testing systems and detecting vulnerabilities compared to previous versions.

According to Oman News Agency, the model is currently accessible through a programme called Trusted Access for Cyber (TAC), which OpenAI launched in February, targeting specific customers and cybersecurity experts to trial its most advanced model. The launch is part of an accelerating race between OpenAI and Anthropic to develop more sophisticated AI models capable of performing complex tasks, particularly in programming and vulnerability detection, as both companies seek to persuade organisations to adopt their technologies.

Last week, Anthropic announced its Mythos model, designed to detect vulnerabilities across operating systems and web browsers, and limited its use to trusted partners including Amazon, Microsoft, Google, and Apple to enhance the protection of their systems.