Anthropic has unveiled Claude Fable 5, an advanced AI model described as the most powerful it has made widely available to date. This release is made possible by the implementation of new safeguards, addressing previous concerns that a Mythos-class model of this capability would be too risky for public deployment.
According to the company, Fable 5 demonstrates "exceptional performance in software engineering, knowledge work, and vision." Its superiority over existing models becomes particularly evident as tasks increase in length and complexity, showcasing its robust capabilities across diverse applications.
The introduction of Fable 5 marks the inaugural widespread release from Anthropic’s Mythos class of AI models. Previously, this family of models was deemed too hazardous for public access due to its profound capabilities in cybersecurity tasks. Anthropic states that the current release is "made possible by new safeguards that block responses in specific high-risk areas," with the system designed to revert to Claude Opus 4.8—a model lauded for its "honesty" upon its launch last month—should such a high-risk scenario arise.
Anthropic specifically highlighted cybersecurity and biology as two critical domains where these protective safeguards might intervene, given their sensitive nature for advanced AI systems. The company reported that during testing, 95 percent of Fable sessions operated entirely on Fable's own responses, without requiring a fallback to Opus 4.8.
In parallel, Anthropic is also introducing Claude Mythos 5, though specific details remain sparse. A company blog post clarifies that Mythos 5 shares the same foundational model as Fable 5, but with "safeguards lifted in some areas." Currently, access to Mythos 5 appears restricted to organizations participating in Anthropic’s private Project Glasswing initiative, which grants access to Claude Mythos Preview. Existing users within this expanding group will be eligible for a Mythos 5 upgrade, with Anthropic indicating plans to "expand access over time through a more systematic trusted-access program."
The Verge’s request for clarification regarding the models' relationship to Claude Mythos Preview and the rationale behind the "5" numbering, given the apparent absence of prior Fable or Mythos models, did not receive an on-the-record response from Anthropic.
Both new models come with a significantly higher price point compared to Anthropic’s previous flagship, Claude Opus 4.8. Pricing is set at $10 per million input tokens and $50 per million output tokens, effectively doubling the rates for Opus 4.8, though it represents half the cost previously associated with Mythos Preview access.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.