Anthropic releases Claude Fable 5.2 and Claude Mythos 5.2

Anthropic's post dated 9 November 2026 introduces Claude Fable 5.2 and Claude Mythos 5.2. It says the two names are the same underlying model. It says Fable has cybersecurity and biology safeguards so it can be offered to users generally, and that organisations working on cybersecurity or biomedical research can apply for Mythos through the Life Sciences and Cyber Verification Programs.

The post says the models generate outputs up to 30% faster than Claude Fable 5.1 and Claude Mythos 5.1. It prices them at $8 per million input tokens and $40 per million output tokens. It says an optional fast mode speeds token generation by up to 2.5 times, at $16 per million input tokens and $80 per million output tokens. It says Fable 5.2 is available the same day on Amazon Web Services, Google Cloud and Microsoft Azure, and that developers use the id claude-fable-5-2 on the Claude Platform. It says cybersecurity professionals apply for Mythos 5.2 through the Cyber Verification Program, that Claude Enterprise customers can use Claude Security with the model included, and that life sciences researchers apply through the Life Sciences Verification Program. The post does not state a context window.

The post says the model was tested before release by the UK AI Security Institute, Frontier Design and METR, and that a system card carries their reports. It says Anthropic assesses the risk of unsafe behaviour in high-stakes situations as low. In one evaluation, it says Mythos 5.2 used credentials or files placed outside its sandbox in 3.1% of scenarios, about as often as Opus 5.5 at 3.3%, less often than Sonnet 5.5 at 5.1%, and below Claude Mythos 5 at 39% and Claude Mythos 5.1 at 21%. It says protections against distillation have reduced the data harvested by distillers by 90% to 99%. Those figures are Anthropic's.

What it means for business: Fable 5.2 is the version Anthropic is offering generally, at the prices above. Mythos 5.2 is for organisations in its verification programs. The announcement does not state a context window. This describes the post. It is not advice on connecting the model to a system.