# Anthropic launches Message Threads in public beta for Claude

published: 2026-11-03T13:05:26.906Z
topic: infrastructure
source: https://the-decoder.com/anthropics-message-threads-cut-time-to-first-token-by-35-percent-for-long-conversations/
url: https://aiagentnews.io/stories/anthropic-launches-message-threads-in-public-beta-for-claude

Anthropic has released Message Threads as a public beta on its Claude platform, letting clients send only what has changed since the previous message instead of the full conversation history. The company says this cuts time to first token by about 35 percent for requests between 500,000 and one million tokens.

Until now, every API request had to resend the entire conversation history. With Message Threads, only the changes since the last message are transmitted, and the thread stays encrypted in memory between requests. A thread expires after five minutes of inactivity.

Anthropic says coding agents, computer-use agents and workflows that process large numbers of documents or images benefit most. Token usage does not change, because Claude still reads the full conversation history on each request.

The feature works on the Claude platform and through AWS, Microsoft Foundry and Google Cloud. Details are in Anthropic's documentation.

## Why it matters if you build agents

Agents that resend long histories on every turn may see faster first responses, especially at very large context sizes. Since token usage stays the same, the gain is in latency rather than in token count. Code that pauses for more than five minutes should expect the thread to have expired before the next request.