Anthropic Launches Message Threads to Speed Up Claude Chats
Anthropic has released Message Threads in public beta, letting developers send only the newest message to Claude and cut time to first token by about 35%.

Anthropic has launched a new feature called Message Threads for its Claude platform, available in public beta. It is designed to make Claude respond faster during long conversations and complex tasks that involve large amounts of data, and it targets developers who use Claude's tools to build AI agents.
The feature changes how conversation history reaches the model. Instead of resending the entire conversation with every request, a developer can send only the latest message together with the ID of the model's previous response. The system then draws on the earlier context cached on the server, so old messages do not have to travel across the network again.
35% faster to the first token
According to Anthropic, Message Threads can cut time-to-first-token by roughly 35% for large requests of between 500,000 and one million tokens. That matters most for agents that write code over long sessions, control computers, or analyze big documents and images.
The company stresses that API pricing and usage limits do not change, because Claude still processes the full conversation context when it prepares an answer.
Retention rules and availability
Conversations are cached temporarily in server memory. A thread expires five minutes after the last request, after which the developer must send the history again. Anthropic says this data is not written to conventional storage and is deleted entirely within an hour of last use.
The feature is in public beta on the Claude Platform and is also supported by Amazon Web Services, Microsoft Foundry and Anthropic-managed models on Google Cloud. It is enabled by default for most Claude Code users. Accounts under Zero Data Retention policies, or those that must meet HIPAA requirements for health data, cannot use it.
Apps worth a look
Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.
Talk to an AI conversation partner; every sentence comes back corrected and explained in your language. 9 languages, A1 to C2.
Open EDIFACT and X12 messages, see every segment explained, convert to JSON/XML, export PDF and Excel. Runs locally, no upload.
Pick who it is for, the occasion and their personality. The search engine reads product descriptions from local shops and ranks what fits. Free, no sign-up.
More from this day
See all →
RapidClaims Reportedly Eyes $30 Mn Round; CEO Denies Talks

Binance $6.3 Billion Heist Began With a Fake Homebrew on Mac

Canva Lays Off 22% of Staff and Cuts Growth Forecast on AI Costs




