Claude Fable 5

Claude Fable 5 is Anthropic's most capable widely released model, with exceptional performance across software engineering, vision, knowledge work, and scientific research. It maintains focus across millions of tokens and improves its own outputs using persistent, file-based memory, compressing long-running work into far fewer sessions. The model achieves state-of-the-art vision performance, extracting precise numbers from detailed scientific figures and rebuilding a web app's source code from screenshots alone. Fable 5 posts top scores on senior-level reasoning benchmarks and, in its research-grade form, consistently produces novel, compelling scientific hypotheses when paired with domain expertise. With a 1M token context window, the model excels at extensive codebases, lengthy contracts, and comprehensive research documentation. Built-in classifiers detect a narrow set of high-risk requests and hand them off to a more constrained model, a path that on average affects fewer than 5% of sessions.