Anthropic unveils Claude 5.1 models with steep cache read price cut
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1, which share the same underlying model but differ in access: Fable is generally available with production safeguards, while Mythos is restricted to vetted cybersecurity and life-sciences organizations. The new release includes a 75% reduction in cost for Fable cache reads, a key benefit for enterprise buyers. The launch marks a significant update to Anthropic's large language model lineup.
Anthropic's latest release introduces two model variants built on a shared foundation. Claude Fable 5.1 is openly available with production safeguards for general commercial use, while Claude Mythos 5.1 is limited to vetted organizations in cybersecurity and life-sciences fields, reflecting a deliberate tiered-access strategy.
The most notable commercial change is a 75% reduction in cache read pricing for Fable, directly targeting a recurring cost concern for enterprises that run frequent, repeated queries. By lowering this expense, Anthropic aims to make its large language models more attractive for high-volume deployments, marking a significant competitive move in the AI model market.
The steep price cut could lower financial barriers for businesses integrating AI into routine operations, potentially accelerating adoption among enterprises that rely on frequent model interactions. The restricted Mythos tier may shape how sensitive sectors like cybersecurity and biomedical research approach AI, though