Anthropic cuts Fable 5.2 pricing by 20%. Mythos shares its core but stays limited to vetted organizations

Anthropic released Claude Fable 5.2 at $8 per million input tokens and $40 per million output tokens, 20% below Fable 5.1. Claude Mythos 5.2 uses the same model with fewer cyber and biology restrictions, but access is reserved for vetted organizations.

Fable and Mythos share a model but differ in safeguards

Anthropic claims output up to 30% faster than 5.1. Fast mode promises up to 2.5 times faster generation at $16 and $80 per million tokens. Fable is available through AWS, Google Cloud and Azure. Mythos is limited to Cyber and Life Sciences Verification Programs.

The most expensive model is meant to advise and orchestrate

Anthropic recommends starting most coding work with Opus 5.5. Fable is positioned as an Advisor for hard problems or an Orchestrator that delegates simpler tasks to cheaper models. The discount aims to make a frontier model usable as a layer in multi-agent systems.

The largest scientific claims rely mainly on Anthropic tests

The company describes two open black-hole physics problems, a Fermat formalization in 30 hours and an operating system built by more than 1,000 subagents over eight days. These are company-run experiments, not independent reproductions. Comparisons also sometimes vary effort, safeguards or harnesses.

Cost per solved task will matter more than a benchmark win

Teams need to measure total completed-task cost, human corrections and long-run stability. For Mythos, the number of organizations that pass vetting and the fit between monitoring and data protection will matter just as much.

Lilith's verdict

Anthropic has seated Fable as an expensive specialist that cheaper models consult on hard cases. It pays off when the invoice shows fewer consultations and more problems carried to completion.