Anthropic's Message Threads cut time to first token by 35% on long conversations

The Decoder · rss · 2026-11-03

Anthropic released Message Threads for the Messages API. Instead of resending the entire conversation with every turn, agents now only transmit what's changed. For requests between 500K and 1M tokens, time to first token drops by about 35%.

Original post →

More from coding & agent

coding & agent channel →