• Home
  • Search
  • AI Services
  • Models
  • Writing
  • Multimedia Creation
  • Translation & Others

Anthropic Claude Fable 5.2

Anthropic Claude Fable 5.2, the newest top-tier Claude model, with stronger coding, reasoning, and chart reading, clearer answers, and faster output.

Anthropic Claude Fable 5.2, released in November 2026, succeeds Fable 5.1 as the most capable tier of the Claude family, positioned for the most demanding reasoning and long-horizon agentic work. Against Fable 5.1 it raises an agentic coding benchmark from about 32 to 49 percent, scientific research terminal tasks from about 53 to 75 percent, and chart reading from about 45 to 73 percent, with clear gains in business workflow automation, computer use, and knowledge work; it communicates more clearly, generates output up to 30 percent faster, and needs fewer tokens for hard tasks at medium effort. Its knowledge cutoff is June 2026. Supports very long context for whole-codebase or multi-document analysis, plus a reasoning-effort dropdown (five levels, default high), an option to show the thinking process, web search, and image and document upload; extended thinking is always on and cannot be switched off. It sits in the top Claude cost tier, below Fable 5.1. Reserve it for the hardest work such as large refactors, complex architecture design, and deep cross-document research; use Opus 5.5 or Sonnet 5.5 for daily work. If you need to search the web for information, please turn on the "Web Search" switch. Prompt example 1: (Upload a Python service's main modules as txt) Plan a migration of this service to Rust: map the modules, list phased steps, a test strategy, and a rollback plan. Prompt example 2: (Upload a revenue trend chart with several lines as png) Read the quarterly values of each product line into a table, name the fastest-growing and steepest-falling lines, and explain your reading. Prompt example 3: Formally prove in Lean 4 that the sum of any two even numbers is even, explain what each tactic does, then show how to extend the proof to any number of even numbers. Prompt example 4: Design a block diagram for a battery-powered indoor air quality sensor, justify the main components, estimate standby and measuring power draw, and list board layout cautions. Prompt example 5: (Upload a company's financial statements as PDF) Build a forecast input table of three years of revenue, gross profit, and operating expenses, check that regional figures sum to the totals. Prompt example 6: Design a market research workflow run by one lead agent and five subagents: define each role, the handoff formats, acceptance checks, and how the lead merges and verifies the results. Prompt example 7: Write SQL that computes monthly net revenue and repeat purchase rate per region from orders, customers, and returns tables, explain how to avoid double counting, and include test data to verify it. Prompt example 8: Derive Planck's law of blackbody radiation, show which classical result it reduces to at long and at short wavelengths, and write Python code that plots the curves for three temperatures. Prompt example 9: Plan the motions for a two-arm robot to unscrew a bottle cap in four stages (perception, grasp, applying torque, confirmation), with likely failures and corrections for each stage. Prompt example 10: (Upload three contract drafts that reference each other as docx) Find clauses that contradict each other or define terms inconsistently, explain each risk, and propose revised wording. 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). Each uploaded file must be smaller than 100 MB; for a larger file, please use the GT Agent desktop software. 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 Pricing: Very high Fast Mode: uses 2x credits Continuous conversation: Supported AI can make mistakes. Please verify important information.