Anthropic releases Claude Fable 5.2 and Mythos 5.2

Anthropic has updated its most powerful Claude models: Fable 5.2 and Mythos 5.2. They are the same model in two versions. Fable is available to everyone, while Mythos goes to organizations that have passed a special vetting. The new version responds up to 30% faster than the previous one and is cheaper: $8 per million input tokens and $40 per million output tokens. There is also a fast mode: 2.5 times faster, but twice as expensive. On the FrontierCode 2.0 benchmark, Fable 5.2 scored 49.1%, against 45.5% for Opus 5.5 and 42.8% for GPT-6 Astra. Anthropic also claims that during testing the model formalized the proof of Fermat's Last Theorem in 30 hours. For most tasks, the company still recommends starting with Opus 5.5.