Anthropic Says Message Threads Cut Time to First Token 35% on Long Requests
Summary
Anthropic releases message threads in public beta on the Claude Platform, letting agents send only new messages per turn instead of resending full history. Anthropic says requests of 500k to 1M tokens cut median time to first token by about 35%. Threads expire five minutes after the last request, and token usage is unchanged.
Key Points
- Anthropic turns message threads on by default for most Claude Code users.
- Anthropic keeps each thread encrypted in memory and writes nothing to disk between turns.
- Message threads also runs on Claude Platform on AWS, Microsoft Foundry, and Google Cloud with Anthropic management.