Articles by tag: claude-mythos

Anthropic AI finds vulnerabilities in OS and browsers

Claude Mythos Preview detects zero-day and creates exploits. Reaction of the Bank of England and banks to new cybersecurity risks. Learn about the consequences for the financial sector.

Claude Mythos passed 32-step cyberattack TLO

Claude Mythos Preview autonomously captured the network in TLO simulation for the first time in hours, not 20 by an expert. CTF results, limitations and AISI plans for tests. For developers and security specialists.

Claude Mythos: powerful cybermodel alarmed FCA and Federal Reserve

Claude Mythos from Anthropic independently searches for vulnerabilities in IT systems. UK and US regulators are holding meetings. Learn details of AISI tests and risks to infrastructure — read analysis for IT specialists.

Claude Mythos found a bug in OpenBSD 27 years

Claude Mythos Preview detected vulnerabilities in OpenBSD, FFmpeg, Linux. Benchmarks, partners, funding. Learn how AI changes security audits for developers.

Claude Mythos: Leak and Cyber Power of Anthropic

Anthropic leaked data about Claude Mythos — top model focused on cybersecurity. Comparison with Opus, attack risks, and limited release. Read details for developers.

System Card Mythos: RSP 3.0 and linear probes

Analysis of System Card Claude Mythos: transition to Risk Reports, analysis of linear probes activations, USAMO benchmarks 97.6%. Practical conclusions for AI products. Study the changes in risk assessment.