Anthropic launches Message Threads to speed up Claude
Anthropic's new beta feature sends only what changed since the last message, cutting wait time for very long conversations, according to The Decoder.
The facts
- Anthropic released Message Threads as a public beta on its Claude platform, according to The Decoder.
- Until now, every API request had to resend the whole conversation history. The new feature sends only what changed since the last message.
- Anthropic says time to first token drops by about 35 percent for requests between 500,000 and one million tokens.
- The thread stays encrypted in memory and expires after five minutes of inactivity. Token usage does not change, and the feature also works through AWS, Microsoft Foundry and Google Cloud.
Case notes · Danny's take
Why it matters
Speed is becoming as important as raw model quality. Agents that run for hours carry huge histories, and resending all of it on every call is wasteful. Notice what did not change: token usage stays the same, so this fixes waiting time, not the bill. The cost story is still separate.
For builders
If you build coding or computer-use agents, or pipelines that chew through documents and images, test this in the beta. Measure your own time to first token, and design around the five-minute expiry so idle sessions do not break. Do not budget for savings, because token counts stay the same.
Not financial advice. Do your own investigation.
This story was summarised with AI from The Decoder's reporting and checked against our editorial policy. Danny's take reflects his views. Disclaimer.
Danny Walter
Founder of Moon Republic, Volt Terminal and Moon Republic Academy. Author of The Market Is a Crime Scene and After the Applause.