Anthropic new feature cuts AI response times by around 35 per cent for heavy tasks.

Anthropic rolls out Message Threads in public beta to make Claude AI much faster

Anthropic's new Message Threads beta makes Claude AI much faster for big tasks without changing the price.

by · India Today

In Short

  • Anthropic has launched Message Threads in public beta for the Claude platform.
  • The new feature cuts AI response times by around 35 percent for heavy tasks.
  • This update is already turned on by default for most Claude Code users.

Anthropic has introduced a new feature called "Message Threads" for its Claude AI platform. Announced in a Monday blog post, the feature is currently in its public beta phase. It is designed to make the AI respond much faster during long and heavy conversation sessions.

How Message Threads work

Usually, when developers build AI agents (like coding assistants), they have to upload the entire chat history to the AI every time they ask a new question. As the chat gets longer, sending this massive block of data takes a lot of time and slows things down.

With Message Threads, developers can skip this step. They now only need to send their newest message along with the ID of the AI's previous response. Because the system remembers the rest of the chat for a short period, it doesn't need to receive all the old text again over the network.

Anthropic notes that this cuts the "time-to-first-token", the time it takes for Claude to start typing its reply, by roughly 35 per cent for massive requests containing 500,000 to 1 million tokens. This speed boost is a big deal for developers running AI agents for complex jobs, like long coding sessions, computer-use agents, or processing heavy documents and images.

Pricing and Privacy Rules

Even with the faster speeds, Anthropic has confirmed that API pricing and rate limits will stay exactly the same. Because Claude still reads the full conversation on its end to give you a good answer, token usage does not change.

However, there is a strict time limit. Threads are only stored temporarily in the server's memory and will expire five minutes after your last request. If a thread expires, the developer will have to resend the full conversation to start a new one. For security, these threads are never saved to standard disk storage and are completely deleted within an hour of their last use.

Who gets it?

The Message Threads beta is available right now on the native Claude Platform. It is also available across Amazon Web Services (AWS), Microsoft Foundry, and Anthropic-managed models on Google Cloud. For most Claude Code users, the company confirmed that the feature is already turned on by default.

There is just one catch: accounts that use strict Zero Data Retention (ZDR) policies or require HIPAA compliance are not eligible to use the Message Threads feature.

- Ends