Anthropic Claude Fable 5, the most powerful and most intelligent Claude model, a new tier above Opus. State-of-the-art deep reasoning, long-horizon autonomous work, and code understanding and generation, surpassing Opus 4.8 on the hardest analyses, large refactors, and complex system design. Supports very long context for whole-codebase or multi-document analysis. Premium pricing (about twice Opus 4.8); reserve it for the most demanding tasks and use Opus or Sonnet for daily work. If you need to search the web for information, please turn on the "Web Search" switch. Prompt example 1: Design an architecture handling one million orders a day, explaining trade-offs in database sharding, caching, message queues, and failover, with a text diagram. Prompt example 2: (Upload three industry white papers from different years as PDF) Compare how their main arguments evolved, identify contradictions, and write a research summary. Prompt example 3: Prove that the product of any three consecutive integers is a multiple of 6, then prove the product of n consecutive integers is a multiple of n!, step by step. Prompt example 4: Design the sync mechanism for a document many people edit at once, compare OT and CRDT on conflict handling, offline edits, and performance, and justify your final choice. Prompt example 5: (Upload about 5,000 lines of legacy COBOL payroll code as txt) Explain the overall flow and each section's business rules as a spec, and flag parts that are unclear or possibly wrong. Prompt example 6: Two coffee shops compete on the same street. Use game theory to analyze their opening hours and target customer strategies, find the Nash equilibrium, and explain when cooperation beats competition. Prompt example 7: Derive the period of a pendulum at large swing angles, explain why the small-angle approximation breaks down, and estimate the percentage difference at a 60-degree amplitude. Prompt example 8: (Upload a week of server error logs as txt) Find causal links between the errors, infer the most likely root cause, reconstruct the incident timeline, and propose prevention measures. Prompt example 9: Do threat modeling for a medical appointment app: list attack scenarios by STRIDE category with impact and countermeasures, then rank what to address first. Prompt example 10: Plan a zero-downtime migration of a 500-million-row member table to a new schema, covering dual writes, data validation, traffic cutover, and rollback steps. Maximum context window (input and output combined): about 1,000,000 tokens. Maximum output per reply: about 64,000 tokens (model reasoning, if any, counts toward this limit). Maximum total upload size per request: 512MB (all files and message content combined). Accepted document formats: pdf, docx, xlsx, pptx, txt, csv (file content is provided to the model as extracted text). Accepted image formats: gif, jpeg, jpg, png, webp. Animated GIF images are analyzed using only the first frame. Maximum number of images: 600. When uploading more than 20 images, each side must be at most 2,000 pixels. Maximum size per image: 10MB (base64-encoded). Maximum total image size per message: 20MB (text and history share the same request size limit; uploads close to the limit may still be rejected). Image dimensions: 200 to 8,000 pixels per side. Output content formats: Text/Image Pricing: Very high Continuous conversation: Supported AI can make mistakes. Please verify important information.