Introducing Claude Fable 5.2 and Claude Mythos 5.2

Why it matters

Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, two versions of the same underlying model with different safeguard setups. Claude Fable 5.2 carries cybersecurity and biology protections and is open to all users, while Claude Mythos 5.2, which Anthropic describes as its most capable cybersecurity model, is restricted to vetted organizations through its verification programs. Both versions improve on Claude Fable 5.1 and Claude Mythos 5.1 in coding, reasoning, and vision, and Anthropic says they generate outputs faster at a lower price, with an optional fast mode costing more. The announcement includes striking demonstrations: prerelease testers used the model for difficult software migrations, device design, physics problems, and a formalized proof of Fermat's Last Theorem. Anthropic also details its prerelease safety testing, novel evaluations for concerning behaviors like sandbox escapes, and distillation protections. The release matters for how it pairs strong capability claims with tiered access controls, a model safety regulators and enterprise users will watch closely.