Anthropic Claude Opus 5, the new-generation flagship. It approaches the frontier intelligence of Fable 5 for roughly half the cost, which makes it the best-value high-end option. Against Opus 4.8 it roughly doubles coding performance on Frontier-Bench, scores about three times the next-best model on novel problem solving, and leads clearly on computer use and scientific research; charts and interactive output are noticeably better too. Its most praised trait is thoroughness: it verifies its own work, iterates carefully, and catches logical errors before acting. 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. Reserve it for the hardest work such as large refactors, complex architecture design, and deep research; use Sonnet 5 or Haiku 5.5 for light everyday chat. If you need to search the web for information, please turn on the "Web Search" switch. Prompt example 1: Build an interactive mortgage calculator in a single HTML file: enter loan amount, annual rate, and term, and instantly show the monthly payment and a total cost chart. Prompt example 2: Explain why temperature and pH affect enzyme activity, and design a lab experiment high school students can do to verify it, with materials, steps, and expected results. Prompt example 3: Find the logical fallacy in this argument and explain the correct reasoning: If it rains, the ground gets wet. The ground is wet now, so it must have rained. Prompt example 4: Explain the esterification mechanism for making ethyl acetate step by step, describing the electron movement, and why concentrated sulfuric acid and removing the water formed raise the yield. Prompt example 5: Explain protein structure from primary to quaternary, use hemoglobin to show how structural changes affect oxygen binding, and end with five concept-check questions with answers. Prompt example 6: Plot this as a dual-axis chart with temperature as a line and rainfall as bars, then describe the seasonal pattern: Jan 16°C, 80 mm; Apr 23°C, 150 mm; Jul 30°C, 250 mm; Oct 25°C, 120 mm. Prompt example 7: Write a JavaScript function that spells out check amounts in words, e.g. 1050.25 becomes One thousand fifty and 25/100 dollars, then list tricky edge cases yourself and verify each one. Prompt example 8: (Upload a csv of daily store visitors) Find weekly patterns and outlier days, build a model to forecast next month's visitors, and backtest it on the previous three months. Prompt example 9: Three roommates share an apartment with rooms of different sizes and light. Design a fair way to split the rent, explain the math, and check whether anyone would still feel shortchanged. Prompt example 10: A bag holds 3 red and 5 blue balls. Each drawn ball is returned with one more of the same color. Derive the chance that the 10th draw is red, then verify it with a simulation. 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: High Fast Mode: uses 2x credits Continuous conversation: Supported AI can make mistakes. Please verify important information.