Anthropic is boosting the capacity of its Claude Sonnet 4 AI model, giving enterprise API customers the ability to send up to 1 million tokens in a single prompt. This expansion, aimed at attracting more developers, allows the model to handle inputs of around 750,000 words—equivalent to more than the entire Lord of the Rings trilogy—or roughly 75,000 lines of code. The update increases Claude’s previous 200,000-token limit by a factor of five and surpasses the 400,000-token context window available on OpenAI’s GPT-5.
The enhanced context window is also available through Anthropic’s cloud partners, including Amazon Bedrock and Google Cloud’s Vertex AI. By offering such a high capacity, Anthropic hopes to strengthen its position as a preferred AI coding tool for enterprise customers.
Position in the AI Coding Market
Anthropic has built a strong enterprise base, selling Claude to coding platforms such as Microsoft’s GitHub Copilot, Windsurf, and Anysphere’s Cursor. However, GPT-5’s pricing and coding performance could challenge this dominance. Cursor, for instance, now uses GPT-5 as its default model for new users.
Despite this, Anthropic’s leadership remains confident. Brad Abrams, product lead for Claude, said the increased context window will greatly benefit AI coding platforms, especially for long agentic coding tasks where the AI works autonomously over extended periods. He added that the company is satisfied with API growth and is not overly concerned about GPT-5’s traction.
Why Context Size Matters
In software engineering, larger context windows allow AI to see more of the project at once, enabling better problem-solving and feature development. Anthropic claims it has focused not only on expanding the raw size of the context window but also on improving the “effective context window,” ensuring the model can meaningfully interpret more of the provided input.
Competitors are also pushing boundaries—Google offers a 2 million token window for Gemini 2.5 Pro, and Meta advertises 10 million tokens for Llama 4 Scout. However, research indicates that beyond a certain point, models may struggle to process such vast inputs effectively.
Anthropic will charge more for prompts exceeding 200,000 tokens. API rates for long inputs are set at $6 per million input tokens and $22.50 per million output tokens, up from $3 and $15 respectively for smaller inputs.
What The Author Thinks
The race for the largest context window is becoming as much a marketing battle as a technical one. While bigger numbers sound impressive, the real advantage will come from how effectively models can use that additional space. Anthropic’s focus on “effective” context could give it an edge, but sustained dominance will depend on pricing, developer trust, and consistent real-world performance—factors that raw token capacity alone can’t guarantee.
Featured image credit: Fortune Brainstorm Tech via Flickr
For more stories like it, click the +Follow button at the top of this page to follow us.