Anthropic has released the AI model Claude Fable 5 to the public and launched Claude Mythos 5 for a select group of Project Glasswing partners.
Both models are built on the same foundational architecture. The key difference is that Mythos 5 operates without some of the standard safety restrictions.
Claude Mythos 5 was deployed as part of Project Glasswing, an initiative in collaboration with the US government. The model is an update to Claude Mythos Preview, which was made available in April to a limited group of cybersecurity experts and critical software infrastructure providers.
In the future, Anthropic plans to grant access to Mythos 5 to select biology researchers and participants in the trusted access program.
The Fable 5 model is available to all users through the Claude API and consumption-based Enterprise plans. In the Pro, Max, Team, and seat-based Enterprise subscriptions, its use requires no additional payment until June 22, 2026. From June 23, usage credits will be needed unless the company extends the grace period.
Fable 5 includes a mechanism for redirecting requests related to cybersecurity, biology, chemistry, and model distillation to the more powerful Claude Opus 4.8. According to Anthropic, this fallback occurs in less than 5% of sessions on average.
The cost of using both models is $10 per 1 million input tokens and $50 per 1 million output tokens.
In May, Anthropic released Claude Opus 4.8 and separately introduced the dynamic workflows feature for Claude Code.