Anthropic Launches Claude Fable 5.1 and Mythos 5.1
Original: Claude Fable 5.1 and Claude Mythos 5.1
Why This Matters
Fable 5.1's pricing cuts and new privacy architecture signal Anthropic's push to expand enterprise AI adoption at scale.
Anthropic announced Claude Fable 5.1 and Claude Mythos 5.1 in September 2026, positioning them as the world's most advanced models for coding and knowledge work. Fable 5.1 costs roughly 25% less than Fable 5, with savings up to 45% for agentic workloads.
Anthropic has introduced Claude Fable 5.1 and Claude Mythos 5.1, describing them as the most advanced models for coding, knowledge work, and long-running agentic tasks. The two models share the same underlying architecture but differ in their safeguard levels. Fable 5.1 is generally available, while Mythos 5.1 is accessible only through trusted access programs, with safeguards tailored for cybersecurity and life sciences use cases.
On pricing, Fable 5.1 is estimated to cost approximately 25% less than its predecessor for typical token-billed workloads, primarily through reduced cache-read pricing. For highly agentic workflows, cost reductions can reach up to 45%.
A new Enterprise Frontier Safeguards (EFS) system addresses data privacy concerns by storing data in cloud infrastructure controlled entirely by the customer rather than Anthropic, effectively equivalent to a zero data retention policy. EFS will roll out to enterprise customers in phases starting later in fall 2026. Eligible customers can use Fable 5.1 with zero data retention until EFS is available.
Safeguard improvements include a 60% reduction in false positives for cybersecurity-related content. Fable 5.1 can now assist in discovering software vulnerabilities, though not in developing exploits. For biology, Anthropic established an access program developed in partnership with the US government to enable scientists to use Mythos 5.1's advanced biology capabilities, with enrollment expected to open soon.
Performance benchmarks show Fable 5.1 outperforming Fable 5 across agentic scientific research, terminal coding, multidisciplinary reasoning, and agentic coding tasks.