🔥OpenAI will release a cybersecurity model on par with Claude Mythos

According to Axios citing sources familiar with the matter, OpenAI is finalizing a cybersecurity capability model at the same level as Anthropic's Claude Mythos, planning to restrict its release to a select few companies through its "Trusted Access for Cyber" program. This means that the two leading AI labs have almost simultaneously reached the same conclusion: the cybersecurity capabilities of the most advanced models are so strong that they cannot be publicly disclosed directly; they must first be used by defenders. Anthropic's Mythos security assessment report released today shows that during testing, Mythos autonomously designed multi-step exploit chains to bypass restricted network access, and the attack details…