What changed
Anthropic released Message Threads as a public beta on its Claude platform. The feature reduces time to first token by about 35% for requests between 500,000 and one million tokens. Unlike previous methods, Message Threads only transmits changes since the last message, rather than resending the entire conversation history.
What this means in practice
The update benefits coding agents, computer-use agents, and workflows processing large volumes of documents or images. These use cases often involve long conversations with extensive token counts, making the efficiency gains particularly relevant. The thread remains encrypted in memory between requests and expires after five minutes of inactivity. Importantly, token usage does not change, as Claude still reads the full conversation history.
Limitations and availability
The feature is available on the Claude platform and through AWS, Microsoft Foundry, and Google Cloud. Details are provided in the documentation. The supplied announcement does not specify exact hardware requirements or licensing conditions for the feature.
