Anthropic has increased Claude’s API request size limit from 32 MB to 128 MB, allowing developers to send four times more data in a single request. The update is now live at no additional cost and is designed to support larger PDFs, image-heavy tasks and extended computer-use sessions.
What Has Changed in Claude’s API?
The previous 32 MB request limit created challenges for applications processing large documents, multiple images and screenshot-heavy workflows.
With the new 128 MB limit, developers can send significantly larger payloads directly to Claude through Anthropic’s API.
Why Does the 128 MB Limit Matter for Developers?
The increase supports several practical use cases:
- Large PDF processing: Handle bigger documents and image-heavy files.
- Computer-use automation: Support longer workflows involving multiple screenshots.
- AI applications: Process larger amounts of information in fewer requests.
- Developer productivity: Reduce issues caused by request-size restrictions.
The update is available without a price increase, and developers do not need to make code changes to benefit from the new limit.
Is the New Claude Limit Available Across All Cloud Platforms?
Not yet. The 128 MB limit currently applies to Anthropic’s own API.
Other platforms retain separate request limits:
- Anthropic API: 128 MB
- Amazon Bedrock: 20 MB
- Google Cloud: 30 MB
Anthropic has also introduced server-side conversation history support, allowing applications to send only the latest messages instead of repeatedly transmitting the entire conversation history.
What Does This Mean for Businesses?
Larger API requests can make AI-powered applications more practical for businesses handling complex documents, customer records, visual data and automated workflows.
For companies building AI agents, document-processing systems and computer-use applications, higher request limits can reduce technical constraints and simplify application design.
However, developers should still consider context-window limits, processing costs and application performance when handling large inputs.
The bigger takeaway: AI infrastructure is evolving beyond smarter models. API capacity, data handling and conversation management are becoming equally important for building reliable AI applications.


