Introducing Claude Fable 5.1 and Claude Mythos 5.1

gist

Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 as the same model behind different safeguards. Fable 5.1 is generally available as claude-fable-5-1 at Fable 5's $10/$50 rates, with cache reads cut 75% to $0.25 per million tokens, about 25% cheaper on typical work and up to 45% on agentic work. It scores 52.6% on Terminal-Bench-Science 0.1 versus 24.7% for Fable 5 and 29.0% for Opus 5. Mythos 5.1 is limited to vetted US cyber and life-science programs; Anthropic reports near-50% protein-binder hit rates and GPU-kernel speedups of 2.5x.

ideas

• Same weights, different gates. Fable 5.1 is generally available; Mythos 5.1 is the same model with more permissive cyber and biology safeguards for vetted programs.
• The price cut is cache reads. Input and output stay $10/$50 per million tokens; cache reads drop 75% to $0.25, which Anthropic measures as about 25% cheaper typically and up to 45% on agentic work.
• Science and coding scores jump. Terminal-Bench-Science 0.1 goes from 24.7% (Fable 5) and 29.0% (Opus 5) to 52.6%; Terminal-Bench 4.0 is 55.8% for Fable 5.1 and 60.9% for Mythos 5.1.
• Biology and cyber stay gated. Mythos 5.1 designed high-affinity binders (nearly 50% hit rate across 12 targets) and sped up seven biology models up to 2.5x, but Fable still sends life-sciences R&D and several dual-use cyber tasks to Opus.
• Enterprise privacy is delayed. Enterprise Frontier Safeguards store customer data on the customer's cloud; until that ships later this fall, eligible customers get zero data retention on Fable 5.1.

quotes