Messages API 响应现在包含 usage.output_tokens_details.thinking_tokens,用于报告计费输出令牌中有多少是扩展思考令牌。在流式传输时,该细分仅出现在最终的 message_delta 事件中。无需 beta 标头。
The Messages API response now includes usage.output_tokens_details.thinking_tokens, reporting how many of the billed output tokens were extended thinking. When streaming, the breakdown appears only on the final message_delta event. No beta header is required.
本文内容采集自官方网站,排版和翻译可能与原页面存在差异。
阅读官方全文