The Most Powerful Cyberattack Model Mythos Unsealed? Anthropic Reveals Mythos-Level Model Coming to All Customers in the Next Few Weeks

According to Dynamic Beating monitoring, Anthropic stated at the end of the announcement of Claude Opus 4.8, "We also plan to release a new level of model with higher intelligence than Opus. As part of Project Glasswing, a few organizations are already using Claude Mythos Preview for cybersecurity work. This level of model will require stronger cybersecurity measures before a full release. We are rapidly advancing the development of these safeguards and expect to bring the Mythos-level model to all customers in the coming weeks."

It is worth noting that there is currently a discrepancy in publicly reported information regarding Anthropic's statement about bringing the Mythos-level model to all customers, and the specific model referred to and the definition of "all customers" are still under discussion. On May 23, Anthropic reiterated in the latest update on the software security collaboration plan "Project Glasswing" that, despite internal testing confirming the next-generation security base model Claude Mythos Preview's robust vulnerability discovery and network attack defense capabilities, the company will indefinitely prohibit the public and commercial availability of Mythos-level models until a more robust security alignment mechanism is developed.

---------------------------------

Click the original text link below to join Dynamic Beating · Lark AI News Channel, monitoring global AI trends and news 24/7.