Same badge, new model underneath

Claude Fable 5.1 and Claude Mythos 5.1 are the same model. The difference between them is the safeguards. Fable 5.1 is generally available. Mythos 5.1 ships only through trusted access programs, with looser restrictions built for cybersecurity and life sciences work.

For anyone pointing these models at code, the change that matters is what the safeguards now permit. Fable 5.1 can be used to identify software vulnerabilities, which the previous safeguards blocked. Exploit generation, penetration testing, and binary-based vulnerability scanning are still redirected to Anthropic’s Opus models. You can ask Fable 5.1 where a codebase breaks, but asking it to weaponize what it finds gets routed elsewhere.

The practical result is fewer interruptions. Anthropic says Claude Code users should see around 60% fewer interventions per session from the cyber safeguards compared to the previous safeguards on Fable 5, and that those safeguards block 60% fewer false positives. That counts if you have ever had a benign defensive query flagged and refused.

The safeguards were stress-tested before release. Anthropic ran its own dynamic evaluation, commissioned external testing from two organizations, and had Gray Swan run automated testing. As with Fable 5 and Opus 5, it reports no evidence of a critical-severity jailbreak.

Mythos for vetted defenders

Mythos 5.1 is where the reduced cyber safeguards live. Access runs through the Cyber Verification Program, which today offers reduced-safeguard access to certain Opus- and Sonnet-class models for defensive work and will soon add Mythos-class models. It is currently limited to a set of US organizations, with plans to expand to more domestic and international partners.

With its cybersecurity safeguards off, Anthropic describes Mythos 5.1 as the strongest cyber model it has released, while placing it in the lower category of risk in its Frontier Compliance Framework. Claude Security, the product that scans codebases for vulnerabilities and suggests patches for human review, now runs on Mythos 5.1.

The change that will affect integrations

One item deserves attention if you build on the API. Fable 5.1 ships with strengthened anti-distillation mechanisms. New API accounts, meaning those created from today onward, can no longer manually edit Claude’s prior context in a multi-turn conversation while preserving the transcript of Claude’s prior thinking. This closes a publicly documented technique that distillers used to extract that thinking.

Existing accounts are not affected yet. The change applies to all users with future model releases, and a small number of custom integrations that rely on editing prior context will be affected when it does. If your tooling rewrites assistant turns mid-conversation and keeps the thinking transcript, read Anthropic’s Help Center note and plan the adjustment before a release forces it.

Price and data retention

Cache reads now cost $0.25 per million tokens, down 75%. For typical workloads that works out to roughly 25% less than Fable 5, and up to about 45% for complex coding and highly agentic tasks where cached context dominates. The rest is unchanged: $10 per million input tokens, $50 per million output.

On data retention, Anthropic previewed Enterprise Frontier Safeguards, which store customer data on the customer’s own cloud infrastructure rather than Anthropic’s, with any human review done by the customer by default. It gives enterprises zero-data-retention privacy while keeping misuse detection running. EFS rolls out in phases starting this fall across Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Google’s Agent Platform, and Microsoft Foundry. Until it ships, eligible customers can run Fable 5.1 and Fable 5 with zero data retention.

Fable 5.1 is available today on the Claude API as claude-fable-5-1, and on AWS, Google Cloud, and Azure. It defaults to High effort in Claude Code and Medium in Claude Cowork and on Claude.ai.

---

Contains a referral link.