Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, updating its models for demanding engineering, scientific analysis and long-running agentic work. The company says the two products use the same underlying model but have different access controls: Fable has stricter safeguards and is available to general customers, while Mythos is limited to vetted organizations working in sensitive fields.
The company reports up to 30% faster generation than the 5.1 models, with standard pricing of $8 per million input tokens and $40 per million output tokens. Optional fast mode is priced at $16 and $80 respectively. Fable 5.2 is available on Anthropic's platform and through major cloud providers; restricted Mythos access runs through its verification programs. These are availability and price changes, not a promise that every customer can use the less-restricted variant.
In Anthropic's disclosed evaluations, Fable 5.2 reached 49.1% on FrontierCode 2.0 at Xhigh effort, compared with 32.4% for Fable 5.1. It reports 69.9% on FrontierSWE v2 and substantial gains in robot-control simulations. These results come from the developer's selected evaluation conditions, so cross-provider rankings and production reliability require independent assessment.
Anthropic also describes a 1,000-agent experiment that produced a bootable prototype operating system over eight days, work on hardware prototypes and mathematical formalization. Those are concrete demonstrations with controlled setups, not evidence that the model can autonomously deliver arbitrary production systems or independently establish scientific discoveries.
The release matters because frontier models are moving beyond question answering into long-horizon technical workflows, while providers are creating separate safety-gated channels for capabilities with dual-use implications. Enterprises should evaluate actual task completion, oversight costs and the different restrictions before treating the launch as a universal capability upgrade.