Anthropic has released Claude Fable 5.1 as the generally available counterpart in a paired launch with Claude Mythos 5.1, describing both as its most advanced models for coding and knowledge work and pointing to their research capabilities as an early sign of how AI can contribute to scientific progress. Fable 5.1 and Mythos 5.1 share the same underlying model; the difference lies in safeguards and access, with Mythos 5.1 reserved for Anthropic's trusted access programs focused on cybersecurity and the life sciences. A published system card covers both versions, walking through RSP evaluations, alignment risk, cyber capability assessments, safeguard robustness, and harmlessness, which signals a coordinated safety and capability story rather than a silent capability jump.
Fable 5.1 is positioned as a practical upgrade for teams already running agentic or knowledge-work pipelines. Anthropic highlights roughly 25% lower cost than Fable 5 on typical token-billed workloads, climbing to about 45% savings on highly agentic usage, driven primarily by reductions in cache-read pricing. Alongside the model, the company is rolling out Enterprise Frontier Safeguards (EFS), which aim to deliver zero-data-retention-style privacy while keeping state-of-the-art adversarial protections, and reports meaningful reductions in false-positive blocks, including 60% fewer benign flags in cybersecurity scenarios. Together, these moves frame Fable 5.1 as a fit for organizations that want strong reasoning and tool use on long-context, document-heavy, and agentic workloads, with pricing and privacy tuned for sustained production use.
