Use Claude Sonnet 5 for Massive Document Analysis
Upload massive datasets or entire codebases using the new 1M window.
Switch to the 'claude-sonnet-5' model to utilize a massive 1-million-token context window for processing large-scale documents or complex code repositories.
The Scenario
You need to analyze an entire technical manual or a large codebase that exceeds the 200k limit of previous models.
Before & after
Users had to split large datasets into multiple chunks or use Retrieval-Augmented Generation (RAG) to process large files. Building and managing these chunking pipelines could take hours or even days of engineering time.
Migrate to Claude Sonnet 5 to get a 1,000,000 token context window. In about 3–5 minutes, you can update your API calls to use 'claude-sonnet-5' and leverage introductory pricing.
The Prompt
I am moving [PASTE_PROJECT_DESCRIPTION] to your new model. Use your 1M context window to analyze these files: [ATTACH_LARGE_FILES]. Provide a comprehensive summary based on the entire dataset.
Claude Sonnet 5 has launched with a 1M token context window and 128k output tokens. This allows for processing entire codebases or massive document sets in a single prompt. Note that 'adaptive thinking' is now enabled by default, and manual thinking budgets must be removed.
Source
Claude Platform - Claude Platform Docs"Claude Sonnet 5 supports a 1M token context window, 128k max output tokens, and introductory pricing of $2 / $10 per MTok."