Anthropic introduces cheaper Claude Fable 5.2 and Mythos 5.2

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on 9 November, updating its models for demanding technical work while cutting their usage prices. The announcement targets programmers, researchers and businesses running complex AI tasks. Reuters independently reported the launch and its pricing, while the performance improvements remain claims from Anthropic and its testers.

The two names describe the same underlying model with different safeguards and access conditions. Fable is the more broadly available version, with restrictions intended to prevent misuse in cybersecurity and biology. Direct Mythos access remains limited to organizations approved through Anthropic’s cybersecurity or life-sciences verification programs. Buying access to Fable therefore does not provide unrestricted Mythos capabilities.

Standard pricing is $8 per million input tokens and $40 per million output tokens. An optional fast mode costs twice as much, at $16 and $80 respectively. Anthropic says standard generation is up to 30% faster than the previous 5.1 versions, with fast mode offering up to another 2.5-fold increase in generation speed. These figures describe output speed, not a guarantee that every complete task finishes proportionally sooner.

The company reports gains in coding, reasoning and interpreting visual information. Its demonstrations include software migrations and scientific work, but published benchmarks and selected customer examples cannot establish how reliably the models will perform on every organization’s documents or code. Anthropic’s model documentation still recommends Opus 5.5 as the starting point for most workloads, reserving Fable for problems that need more demanding reasoning or longer autonomous work.

Fable 5.2 is available through Claude’s platforms and cloud partners including Amazon Web Services, Google Cloud and Microsoft Azure. Privacy arrangements also matter for enterprise adoption. Anthropic says its Enterprise Frontier Safeguards system, which keeps monitoring data in customer-controlled infrastructure, is now in private beta. Eligible customers can receive zero data retention; other customers face 30-day retention for safety monitoring.

For users, the practical change is a cheaper route to a more capable specialist model. For businesses, it creates a reason to reassess which tasks need the largest model and which can stay with faster, less expensive alternatives. Actual savings will depend on accuracy, retries and human review, alongside the advertised token price.