What changed
Anthropic released one underlying model through two access paths: safeguarded Claude Fable 5.2 on all supported platforms and less-restricted Claude Mythos 5.2 for verified cyber and life-sciences organizations. The API model is claude-fable-5-2, with 1M-token context, up to 128k output tokens, adaptive thinking and optional fast mode.
Why it matters
The release changes the top end of Anthropic's technical-work stack while lowering standard API prices from the 5.1 generation to $8 per million input tokens and $40 per million output tokens. It also makes safety routing, retention and verification status part of model selection rather than an invisible implementation detail.
What to do
Run a matched private eval against Fable 5.1 and Opus 5.5 at the same effort and tool settings; log cost, latency, blocks, fallbacks and task correctness before changing a production route. Apply separately if a verified Mythos workflow is genuinely needed.
Caveat
Anthropic's capability, speed and safety results are mostly first-party or selected early-tester evidence. Mythos is not generally available, fast mode costs $16/$80 per million tokens, and non-enterprise safety monitoring can retain a 30-day usage window.