Researchers Replicate Anthropic's Mythos AI Vulnerability Findings for Under $30
Security researchers have reproduced Anthropic's Mythos vulnerability findings using off-the-shelf models—GPT-5.4 and Claude Opus 4.6—in an open-source harness, at a cost of less than $30 per scan. The replication raises concerns that the attack surface identified by Anthropic is broader than previously understood and accessible to low-budget adversaries.
Реклама: Article Inline