Key breakthrough in security mechanism, Anthropic will release next-generation Mythos level model within weeks at the earliest According to Beating monitoring, Anthropic confirmed in the official announcement of its latest flagship model Claude Opus 4.8 that the development team has made breakthrough progress in security defense and protection mechanisms. It is expected to launch the next generation Mythos level model with super flagship intelligence to all customers as early as the next few weeks. In the Glasswing software security project report released on May 22nd, the research team emphasized that Mythos level models would not be made available to the public until more effective security mechanisms were developed. As a cutting-edge model with strong network attack and defense capabilities such as end-to-end penetration of shooting ranges, the public release of Mythos level models has long been constrained by the serious attack and defense time gap of "discovering vulnerabilities only takes seconds, patching vulnerabilities takes weeks". To prevent the model from being abused and causing serious harm, the Anthropic team is currently accelerating the development of more powerful security defense mechanisms. At the same time, the official has begun to provide targeted scanning and orchestration kits and other tools to eligible security teams to help the defense side establish asymmetric advantages. [Original link]