Anthropic Announces Message Threads in Public Beta to Reduce Latency for Long-Running Sessions

Anthropic has introduced Message Threads in public beta on the Claude Platform. The new feature is intended to reduce response latency in long-running AI agent sessions. It eliminates the need to resend the entire conversation history with every API request. With Message Threads, developers can share new messages and the ID of the previous response to get responses.  Anthropic confirmed that the new feature is enabled by default for most Claude Code users. 

Message Threads Available in Public Beta on the Claude Platform

In a blog post on Monday, Anthropic announced that Message Threads is now available in public beta on the Claude Platform. It is designed to reduce latency for long-running sessions, and the company says that with message threads, agents no longer need to resend the full conversation history on every turn.

Message Threads feature is available now in public beta on the native Claude Platform. It is also available on the Claude Platform on Amazon Web Services (AWS), Claude in Microsoft Foundry, and on Anthropic-managed Claude models on Google Cloud. Anthropic confirmed that the Claude token pricing and API rate limits will continue to apply

A coding agent might make hundreds of requests in one session, and with the newly launched Message Threads, developers can instead submit only the latest messages added since the previous turn. This reduces time to first token (TTFT).

Anthropic says Message Threads default for most Claude Code users reduces median response latency by around 35 percent for requests containing 500K to 1M tokens.

Developers can enable the feature by adding the thread parameter to Messages API requests. The first request includes the full conversation and starts a thread. Each request after that includes only the new messages and the ID of the previous response. Anthropic says that conversation threads are temporarily maintained in memory between requests. A thread expires five minutes after its last request, and the client then has to resend the full conversation to start a new one.

The company also confirmed that Claude will read the full conversation on every Message Threads turn; therefore, token usage stays the same. Anthropic says Message Threads is particularly useful for Long-running coding agents, computer-use agents, and document- and image-heavy workflows, among others.