返回全部动态

Anthropic 推出 Message Threads,长对话首 token 时间缩短 35%

原标题:Anthropic's Message Threads cut time to first token by 35 percent for long conversations

THE DECODERAPI 更新质量 68

AI 摘要

Anthropic 在 Claude 平台公测 Message Threads 功能,改变此前每次 API 请求需重发完整对话历史的方式,仅传输上次消息以来的增量内容。据 Anthropic 称,在 50 万至 100 万 token 的请求中,该功能将首 token 时间缩短约 35%。编码智能体、计算机使用智能体和处理大量文档或图像的工作流受益最大;线程在请求间以加密形式驻留内存,闲置五分钟后过期,token 用量不变,因为 Claude 仍读取完整对话历史。该功能在 Claude 平台及 AWS、Microsoft Foundry、Google Cloud 上可用。

以上摘要由 AI 生成,可能存在误差。事实请以原文为准。

正文节选

Anthropic's Message Threads cut time to first token by 35 percent for long conversations Anthropic released Message Threads as a public beta on its Claude platform. Until now, every API request had to resend the entire conversation history. Message Threads only transmits what's changed since the last message, cutting time to first token by about 35 percent for requests between 500,000 and one million tokens, according to Anthropic. Coding agents, computer-use agents, and workflows processing lot


发布时间:2026-11-03 20:35
抓取时间:2026-11-03 21:13
来源机构:THE DECODER
阅读原文the-decoder.com