Anthropic has introduced a beta feature for its Claude Messages API called "message threads," designed to optimize long agent loops by keeping conversation history on the server. This new functionality means that subsequent requests only need to send new messages, rather than the entire conversation history. This approach significantly reduces request size and improves the time to first token, with Anthropic reporting a median drop of about 35% for requests between 500k and 1M tokens. While token usage and pricing remain the same, this feature streamlines interactions for developers building complex agent applications. AI
IMPACT Streamlines long agent interactions by reducing request size and improving response times for large context windows.
RANK_REASON This is a product update for an existing API, not a new model release or core research.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →