Claude Fable 5.1 and Mythos 5.1: pricing, access and safeguards explained
Anthropic has released Claude Fable 5.1 for coding, long-running agents and complex knowledge work, alongside Claude Mythos 5.1 for advanced cybersecurity and biology research. The two names do not describe completely separate model foundations: Anthropic says they use the same underlying model, with different safeguards and access boundaries.
Where Fable 5.1 is available
Claude Fable 5.1 is available to Pro, Max, Team and Enterprise subscribers. Developers can use it through the Claude Platform and supported marketplaces, including Amazon Web Services, Google Cloud and Microsoft Foundry. The API model name is claude-fable-5-1.
Anthropic prices the model at $10 per million input tokens and $50 per million output tokens. Cache reads cost $0.25 per million tokens. The company says the lower cache price can reduce typical workload costs by about 25% and highly agentic workloads by as much as 45%, depending on reuse patterns. US-only inference costs 1.1 times the standard input and output price.
What Fable is designed to do
The product is aimed at work that can run for hours across tools and applications: software changes spanning a codebase, review and performance work, research, document-heavy analysis and asynchronous business workflows. It can use vision to interpret diagrams, charts, tables and interface output.
These are product capabilities, not a reason to remove human approval from consequential actions. Long-running agents can compound an early misunderstanding, so organisations still need scoped permissions, checkpoints, logs and rollback procedures.
Why Mythos access is restricted
Anthropic describes Mythos 5.1 as its most capable model for cybersecurity and biology research. Access is limited to vetted cyber defenders and life scientists through trusted-access programmes and is currently confined to selected US organisations. Claude Security now runs on Mythos 5.1.
Fable includes safeguards intended to block or limit risky cyber, biology and chemistry assistance. Anthropic says some flagged cyber requests are routed to Claude Opus 4.8 and some biology requests to Opus 5. API customers must configure the Fallback API for this behaviour.
Data retention and benchmark caveats
Using Fable carries a default 30-day data-retention period for safety monitoring. Anthropic says eligible enterprise customers may use Enterprise Frontier Safeguards with customer-controlled storage and, during the transition, some eligible customers can use Fable with zero data retention. Organisations should verify the contract and deployment mode rather than infer privacy terms from the model name.
Anthropic reports improvements on coding, agent and knowledge-work evaluations. Its own benchmark notes warn that some tasks were routed to other models when safeguards intervened, and that updated OSWorld tasks cannot be compared directly with older releases. Independent testing under equivalent tools, prompts and budgets will be needed for a reliable comparison.
Sources: Claude Fable 5.1 product page; Claude Mythos 5.1 product page.
Official Claude Fable product artwork. Image: Anthropic.
