Anthropic Releases Claude Fable 5.1 and Mythos 5.1

Anthropic has officially released its latest frontier models, Claude Fable 5.1 and Claude Mythos 5.1, which the company confirms are two names for the same underlying architecture. This dual-naming strategy appears aimed at different market segments, with Fable positioned for general developer use and Mythos targeted at enterprise clients with stricter compliance needs. The headline feature of this release is a significant pricing adjustment: a 75% reduction in cache read costs for Fable. This change makes it dramatically cheaper for developers to run applications that involve long context windows or repeated calls with similar prompts. For example, a developer building a code assistant that references a large codebase will now pay a fraction of the previous cost to retrieve that cached context, potentially slashing overall inference bills by 30-50% in high-volume scenarios. Beyond the cost reduction, Anthropic claims that both models deliver enhanced performance on reasoning benchmarks, particularly in areas like mathematical proof and multi-language comprehension. The company has also improved the models' ability to follow nuanced instructions, reducing the likelihood of 'hallucinated' tool calls. This release is a clear signal of Anthropic's intent to remain competitive in the frontier AI space, directly challenging offerings from Google and OpenAI. By coupling state-of-the-art performance with aggressive pricing on cache operations, Anthropic is courting developers who have been priced out of high-frequency AI usage. The move is likely to pressure competitors to rethink their own pricing structures. For current users, the upgrade path is straightforward, with API endpoints remaining backward compatible. As the AI landscape becomes increasingly crowded, Anthropic's focus on cost-efficiency alongside raw capability suggests a mature understanding of what enterprises actually need: powerful models that don't break the bank.