Claude Fable 5.2

Claude Fable 5.2 is the same underlying model as Claude Mythos 5.2, released with safeguards for general use. Anthropic aims it at long-running agent work: it leads Opus 5.5 on Anthropic's agentic coding, business-workflow and computer-use evaluations, and thinking is always on, steered by an effort setting. It reads text and images, holds 1M tokens of context and writes up to 128K tokens per reply. It is cheaper than Fable 5.1 ($8/$40 vs $10/$50) but still Anthropic's most expensive tier, so smaller models are the better fit for high-volume, simple work.