Anthropic Releases Public Mythos-Class Model With Cyber Safeguards

Anthropic has released Claude Fable 5, a Mythos-class model built on the same underlying architecture as Claude Mythos 5 but configured for general public use with added safeguards.

The launch confirms the core of Tuesday’s rumors: Mythos-level capability is now available outside the restricted Project Glasswing program. The public model is not the unrestricted cybersecurity version. Anthropic says Fable 5 automatically falls back to Claude Opus 4.8 when classifiers detect requests tied to sensitive areas including cybersecurity, biology, chemistry and model distillation.

Claude Mythos 5 remains limited to Glasswing partners and select trusted-access users. Anthropic says Mythos 5 uses the same underlying model as Fable 5 but has some safeguards lifted for vetted cyberdefenders and infrastructure providers. Existing Claude Mythos Preview users can upgrade to Mythos 5, while broader access is expected through a trusted-access program.

Fable 5 is available through the Claude API and consumption-based Enterprise plans from launch. Anthropic lists pricing for Fable 5 and Mythos 5 at $10 per million input tokens and $50 per million output tokens.

Public Release Follows Wallet-Approval Panic

The release lands hours after crypto users began revoking wallet approvals amid rumors that Anthropic was preparing a public Mythos launch. That reaction started before the official announcement and reflected the crypto industry’s growing concern over AI-assisted exploit discovery.

The panic was not based on a confirmed wallet attack. It came from Anthropic’s earlier Mythos Preview materials, which showed a model capable of identifying and exploiting serious software vulnerabilities in major operating systems, browsers and open-source software. Project Glasswing was created to give vetted defenders access before similar capabilities become more widely available.

Fable 5 changes the public-access side of that story. Anthropic is not releasing the full unrestricted cyber model to everyone, but it is making a Mythos-class system generally available for coding, research, analysis, vision and long-horizon knowledge work.

Safeguards Split Fable 5 From Mythos 5

The central distinction is safeguards. Fable 5 is the public model with classifiers that route sensitive requests away from the Mythos-class system. Anthropic says more than 95% of sessions do not trigger fallback, meaning most users receive the full Fable 5 experience for ordinary tasks.

Cybersecurity requests face stricter handling. Queries involving exploit development, offensive security, agentic hacking or other high-risk cyber activity are blocked from Fable 5 and handled by Opus 4.8 instead. Biology, chemistry and distillation-related prompts are also covered by the new classifier system.

Mythos 5 is the restricted version for approved users. Anthropic says it has the strongest cybersecurity capabilities of any model in the world and will initially be deployed through Project Glasswing in collaboration with the U.S. government.

Crypto Security Debate Moves Into AI Era

The launch gives crypto developers, auditors and wallet teams a clearer timeline for the AI-security shift. Frontier models are becoming stronger at code review, vulnerability discovery, patch writing and long-running technical work. That can help defenders find flaws faster, but it also raises pressure on projects with weak operational security, old dependencies, exposed admin keys, fragile bridges or neglected smart contracts.

Crypto’s concern is not limited to protocol code. Wallet approvals, front-end compromise, DNS attacks, signing prompts, browser extensions, developer machines, multisig procedures and third-party libraries can all become attack surfaces when AI reduces the time needed to search for weaknesses.

That concern has already reached DeFi debates around smart-contract risk. Recent arguments over whether AI could push DeFi security back toward Bitcoin-style simplicity show how quickly Mythos-class models have become part of the crypto threat model.

Anthropic’s public release does not put unrestricted Mythos cyber capability in every user’s hands. It does put Mythos-class intelligence into general circulation with guardrails, while a stronger version remains available to vetted defenders. For crypto teams, the practical response is already visible: fewer stale approvals, stricter signing policies, stronger multisigs, faster patching and more serious security reviews before attackers get the same level of automation.

The post Anthropic Releases Public Mythos-Class Model With Cyber Safeguards appeared first on Crypto Adventure.