The layout in the field of AI security is accelerating. According to renowned tech journalist Alex Heath, Anthropic is expected to officially launch the public version of its cybersecurity model Mythos tomorrow.

Looking back at the development of this model, Anthropic first introduced Mythos in April of this year, with access limited to specific institutions that underwent strict review. This release for the general market means the model will enter a more extensive application phase. To adapt to the public environment, Anthropic will launch this version under the name "Fable5" and integrate more stringent security measures into the product design.

image.png

In terms of functionality, compared to the previous internal version available only to "Glasswing" (Project Glasswing) partners, the upcoming public version of Mythos will have restricted network operation permissions to ensure effective prevention of potential misuse in a more open ecosystem.

Although targeted adjustments have been made to the permissions, the public version of Mythos still demonstrates strong performance. It is reported that the new model has optimized its ability to handle long-term, multi-step tasks. This efficient execution of complex workflows will make it more versatile in scenarios such as network threat detection and security audits, which require deep logical analysis.

This release is a key move for Anthropic to enrich its security AI product line and also reflects the company's commitment to advancing AI technology toward higher security standards. With the official market launch of Mythos, enterprise users will gain a more practical intelligent assistant to deal with the increasingly severe cybersecurity environment.