Anthropic launches Claude Fable 5.2 and Mythos 5.2

Anthropic introduced Claude Fable 5.2 and Claude Mythos 5.2 on Monday, two versions of the same model. Fable carries strong safeguards in cybersecurity and biology and is open to anyone. Mythos relaxes those restrictions, so it is limited to verified organizations.

API customers pay $8 per million input tokens and $40 per million output tokens for Fable 5.2 at the standard rate. A token is the chunk of text a model uses to count what it reads and writes. Fast mode, which generates up to 2.5 times faster, costs $16 and $80. The announcement does not give regional pricing.

Developers can call the model as `claude-fable-5-2` on Anthropic's platform and on Amazon Web Services, Google Cloud and Microsoft Azure. As a launch promotion, subscribers on an eligible paid plan get one to three Fable resets, and on Pro and standard Team seats the model counts for up to half the weekly usage limit.

Access to Mythos 5.2 goes through the Cyber Verification Program, for cybersecurity professionals, or the Life Sciences Verification Program, for life sciences researchers. Claude Enterprise customers can use Claude Security without going through the first.

Anthropic says the model improves at coding, reasoning and vision. Here are its numbers, set against Fable 5.1 and Opus 5.5:

None of the figures has been tested independently. On CVE Rediscovery, which measures how many known vulnerabilities the model rediscovers, Anthropic says Mythos 5.2 scores 89.6%, up from 78.3% for Mythos 5.1. The company also says results arrive up to 30% faster, and that its anti-distillation protections, the strongest it has published, cut by 90% to 99% the data someone trying to copy the model can extract.

For most coding and knowledge tasks, Anthropic itself suggests starting with Claude Opus 5.5.