ChainThink reports that, according to an official announcement, Anthropic revealed in its Claude Opus 4.8 release notice that it will launch a new Mythos-level model with superior intelligence to Opus.
As part of the Project Glasswing initiative, a small number of organizations are currently using Claude Mythos Preview for cybersecurity-related work.
Anthropic states that models at this level require enhanced cybersecurity protections before general release, and the company is rapidly advancing the development of these safeguards, with Mythos-level models expected to be available to all customers in the coming weeks.
However, current public reports differ on this statement, and the specific model referred to and the definition of “all customers” remain unclear.
On May 23, Anthropic reaffirmed in its Project Glasswing update that, despite internal testing demonstrating Claude Mythos Preview's exceptional capabilities in vulnerability discovery and cyber defense, the company will indefinitely prohibit public and commercial access to Mythos-level models until more effective safety alignment mechanisms are developed.