Claude Fable 5.2 (US)

Claude Fable 5.2 is positioned for demanding reasoning and long-horizon agentic work rather than quick chat completions. The documented use cases are agent sessions that run for hours, multistep deep research, and analysis carried all the way through to a finished document, spreadsheet, or deck — meaning the emphasis is on sustained, multi-stage task completion with real deliverables at the end. It sits at the top of Anthropic's current lineup: in the published comparison table it is priced and positioned above Claude Opus 5.5, Sonnet 5.5, and Haiku 5.5, with a higher default reasoning effort than the Opus and Haiku entries and a newer knowledge horizon than the legacy Fable 5 generation.

Technically, the model uses adaptive thinking that stays always on, with a high default effort setting, and it retains the family's large context window and long maximum output allowance, so long working sessions do not need to be split up frequently. A near-identical sibling, Claude Mythos 5.2, shares these specifications and is offered separately to organizations verified through Anthropic's verification programs for defensive cybersecurity use cases, which indicates the same underlying capability is being targeted at specialised, gated workflows. In practice, Fable 5.2 is a good fit for teams building research agents, long-running automation, and document-generation pipelines where quality of the final artifact matters more than per-token cost or response speed.