SECURITY % min read

OpenAI opens its cybersecurity model to thousands of defenders in race with Anthropic’s Mythos

OpenAI opens its cybersecurity model to thousands of defenders in race with Anthropic’s Mythos
Photo by Pankaj Patel / Unsplash

OpenAI is releasing GPT‑5.4‑Cyber to thousands of verified defenders, expanding its Trusted Access for Cyber program and positioning itself directly against Anthropic’s tightly restricted Mythos model. The new system lowers refusal boundaries, supports binary reverse engineering, and is designed to answer sensitive defensive security queries for authenticated professionals. Unlike Anthropic’s invite‑only approach, OpenAI argues that broad verified access is essential so smaller organizations aren’t outgunned by attackers. The move highlights a growing philosophical split in cybersecurity AI as both companies race to equip defenders with increasingly powerful tools.

Read the full story on The Next Web