Anthropic has launched Claude Fable 5, a Mythos-class AI model that the company says is safe for general use.
The model is now the most capable Claude model available to the public. Anthropic says it performs strongly in software engineering, knowledge work, vision, scientific research, memory, and long-running tasks.
The release is notable because Mythos-level models were previously limited to selected cyberdefenders and infrastructure providers through Project Glasswing.
Safety Guardrails
Anthropic says Fable 5 uses the same underlying model as Claude Mythos 5 but includes safeguards for general users.
If the system detects requests involving cybersecurity, biology and chemistry, or model distillation, the response will be handled by Claude Opus 4.8 instead of Fable 5. Users will be notified when this happens.
Anthropic says the safeguards are intentionally cautious and may block some harmless requests. However, the company says fallback occurs in less than 5% of sessions on average.
Claude Mythos 5
Anthropic is also launching Claude Mythos 5 for a smaller group of approved users.
Mythos 5 uses the same base model as Fable 5 but removes some safeguards in certain areas. It will first be available to Project Glasswing partners in collaboration with the US Government.
Anthropic plans to expand Mythos 5 access later through a trusted-access program for cybersecurity organizations. It also plans a separate biology access program for selected researchers.
Performance Claims
Anthropic says Fable 5 is stronger than its earlier public models on most tested AI benchmarks.
In software engineering, the company said Stripe used Fable 5 to complete a codebase-wide migration in one day across a 50-million-line Ruby codebase. Anthropic said the same work would have taken a team more than two months by hand.
The company also said Fable 5 performs strongly on coding, finance reasoning, chart and table interpretation, scientific vision tasks, and long-context work.
Research Uses
Anthropic says Mythos 5 has also shown strong results in life sciences.
The company said its internal protein design experts used Mythos 5 to speed up parts of the drug design process by about ten times.
It also said Mythos 5 produced molecular biology hypotheses that Anthropic scientists preferred over Opus-class outputs about 80% of the time in blinded comparisons.
Data Retention
Anthropic is introducing a 30-day data retention policy for Fable 5, Mythos 5, and future models with similar or higher capability.
The company says this data will not be used to train new Claude models or for non-safety purposes. It says the policy is meant to help detect jailbreaks, complex misuse attempts, and false positives.
Pricing And Availability
Claude Fable 5 is available today. Developers can access it through the Claude API as claude-fable-5.
Claude Mythos 5 remains restricted to Glasswing partners until the trusted-access program opens. Both models cost $10 per million input tokens and $50 per million output tokens.
Fable 5 is fully available through the Claude API and consumption-based Enterprise plans. For Pro, Max, Team, and seat-based Enterprise users, it is included at no extra cost through June 22.
On June 23, Anthropic plans to remove Fable 5 from those subscription plans, after which users will need usage credits. The company says it may extend the included access period if capacity allows.
