Claude Sonnet 4 expands context window to 1 million tokens, a 5x increase on Anthropic's API
Aug 7, 2025
Key Points
- Anthropic expands Claude Sonnet 4's context window to 1 million tokens on its API, a 5x increase that lets developers process 75,000+ lines of code in a single request.
- The move narrows the gap with OpenAI's o1 and GPT-4 variants on a key competitive capability for enterprise and developer workflows.
- No pricing changes announced, leaving unclear whether Anthropic will charge more for larger context windows or absorb the higher computational cost uniformly.
Summary
Anthropic expanded Claude Sonnet 4's context window to 1 million tokens on its API, up from 200,000 tokens. The model can now process 75,000+ lines of code in a single request, improving its ability to handle large codebases, lengthy documents, and complex multi-file reasoning tasks.
The move matches competitive pressure from OpenAI's o1 and GPT-4 variants, which already support comparable or larger windows. Context window size has become a material differentiator for enterprise and developer use cases. A larger window reduces the need for developers to chunk inputs or manage retrieval-augmented generation systems, lowering integration friction and broadening the range of problems Claude can solve in one API call. For code-heavy workflows such as debugging, refactoring, and architecture review, the jump from 200k to 1M tokens is functionally significant.
The timing suggests Anthropic believes it can offer larger context windows at acceptable margins for API customers. Larger windows typically increase computational demand per request, so a public expansion signals confidence in scaling Claude's inference without degrading latency or cost economics.
No pricing changes were announced with the expansion. It remains unclear whether Anthropic will tier pricing by context window size or absorb the additional cost uniformly across the API tier structure.