Anthropic Unveils 'Claude Mythos', Powerful AI With Major Cyber Implications

Anthropic has introduced Claude Mythos, a new AI model capable of identifying critical vulnerabilities in software at a scale previously unseen. Mythos is positioned as significantly more advanced than Anthropic's existing models, including Haiku, Sonnet, and Opus. This powerful AI is being utilized in Project Glasswing, a collaborative initiative with major tech companies focused on enhancing software security. Mythos' agentic coding and reasoning skills enable it to excel at software coding tasks and quickly discover previously unknown flaws. The model has already identified thousands of zero-day vulnerabilities, some dating back several decades, that other automated tools missed. Anthropic acknowledges the potential for Mythos to be misused for offensive cyber operations due to its advanced capabilities. They are working to deploy Mythos strategically to improve cybersecurity before bad actors can exploit its power. Project Glasswing is a proactive effort to leverage Mythos to strengthen security across various software systems. Anthropic intends to involve other developers, researchers, and governments in this effort. The Claude Mythos Preview is a closed version that will never be released to the general public.