Anthropic Launches Claude Fable 5: First Public Mythos-Tier AI Model

Anthropic today officially launched Claude Fable 5, its first public Mythos-tier model, stating that the model has passed safety assessments and is ready for general-purpose deployment. According to official details, Fable 5 surpasses all previously released Anthropic models in overall capability, delivering "exceptional performance" across software engineering, knowledge work, visual recognition, and scientific research. Notably, queries involving sensitive areas such as cybersecurity, chemistry, and biology will default to Claude Opus 4.8 instead of Fable 5 to strengthen safety controls.

Another major highlight of this release is Mythos 5 — deployed by Anthropic through the Project Glasswing initiative as an upgrade to the Claude Mythos Preview. Anthropic claims Mythos 5 possesses the world’s strongest cybersecurity capabilities and plans to gradually expand access through a broader "trusted access program." Current Mythos Preview testers, including Apple, can already use Mythos 5 directly. On the same day, Anthropic also announced the release of an updated Claude Opus 4.8, further enriching its model lineup.

From an industry perspective, Fable 5 marks a key step in Anthropic’s commercialization of high-end AI models — leveraging the "Mythos" tier to differentiate capabilities while controlling risk through safety mechanisms, balancing performance with trusted AI commitments. As Project Glasswing advances, Mythos 5 — with its strongest safety capabilities — is poised to create a competitive edge in the enterprise cybersecurity market.