Claude Fable 5.2

Claude Fable 5.2 arrived on November 9, 2026 as the successor to Claude Fable 5.1. It is the same underlying model as Claude Mythos 5.2, which Anthropic offers only to organizations verified through its cyber and life sciences programs; Fable 5.2 carries cyber and biology safeguards that let it be offered to everyone. It keeps the 1 million token context window and 128K output ceiling, takes text and images, and thinks adaptively on every request, steered by five effort levels from low to max with high as the default. Anthropic positions it for work that runs for hours: long agent sessions, multistep deep research, and analysis carried through to a finished document, spreadsheet or deck. It reports output up to 30% faster than Fable 5.1, stronger chart and image reading, and steadier coordination when the model orchestrates subagents or advises another model. A few behaviours matter when moving to it. Forced tool choice returns an error, so schema-shaped output relies on strict tool use or JSON outputs. Thinking blocks it produces are read only by Fable 5.2 and Mythos 5.2, so a conversation moved to another model loses the earlier reasoning, and editing earlier turns invalidates those blocks. Anthropic recommends starting at the default effort and stepping up to xhigh only where a gain is measured.