Event ·Official first-hand

OpenAI GPT-Live API 新增 Standby 配置:闲置时通话可降费率保持不断线

OpenAI 在 GPT-Live API 中推出 Standby 配置,覆盖 GPT-Live-1 和 GPT-Live-1.1 两个版本。开发者可将空闲超时设为 5-60 秒,用户静默超过该时长后,模型会临时暂停,但 WebRTC 通话保持连接。 费用机制是关键差异:Standby 期间费率从标准 $0.05/分钟降至 $0.02/分钟,降幅 60%。一旦有人再次开口,模型恢复响应,会话不会中断,用户无需重新发起通话。 对语音类产品团队来说,这解决了语音会话中用户长时间停顿导致的高额空转计费问题。此前开发者要么让连接超时断开、牺牲体验,要么持续按 $0.05/分钟付费,Standby 在两者间提供了折中方案。

Current read

做语音产品的话这个挺实用:用户不说话时,WebRTC 连接不断但费率从 $0.05 降到 $0.02 一分钟。以前要么断线要么白烧钱,现在 5-60 秒可设空闲阈值,开口即恢复。接 GPT-Live-1 和 1.1 的开发者都该看。

11 sources41° AI heatLast updated 2026/11/19 20:00:39

Evidence

11 sources
Primary source@OpenAIDevs
View source
Related source 1TestingCatalog
View source
Related source 2Databricks
View source
Related source 4Greg Brockman
View source
Related source 6向阳乔木
View source
Related source 7Perplexity
View source
Related source 8OpenRouter
View source
Related source 10openclaw
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.