OpenAI Launches GPT-Live-1.1 API: Supports Real-Time Two-Way Conversations, Allowing Agents to Handle Complex Tasks While ChattingOpenAI 开放 GPT-Live-1.1 API:支持实时双向对话,智能体可边聊天边处理复杂任务
According to IT Home on 20 November, OpenAI announced today that it is making its GPT-Live-1.1 API voice model available to developers, allowing third-party developers to integrate it into their own applications to build AI that can listen, respond, and perform complex tasks in real time…
IT之家 11 月 20 日消息,OpenAI 今日宣布向开发者开放 GPT-Live-1.1 API 语音模型,允许第三方开发者将其集成到自己的应用程序中,构建能够实时聆听、回应并执行复杂任务的 A…

Key Points要点速览
According to a report from IT Home on 20 November, OpenAI announced today that it is making its GPT-Live-1.1 API voice model available to developers, allowing third-party developers to integrate it into their own applications to build AI voice agents capable of listening, responding, and performing complex tasks in real time. According to OpenAI, GPT-Live-1.1 is an upgraded version of GPT-Live-1, featuring full-duplex voice interaction capabilities that allow it to listen and speak simultaneously, without waiting for the user to finish a sentence before starting to process it. The model can recognise pauses, interruptions, and changes in speech rate as the user speaks, and decide in real time whether to continue listening or to respond. IT Home Note: GPT-Live-1.1 has been the default voice model for ChatGPT Plus and higher-tier subscribers since 15 October 2026. This API release further extends the relevant capabilities to third-party developers. In terms of specific capabilities, OpenAI states that GPT-Live-1.1 has improved multilingual speech processing, instruction-following capabilities, and speech fidelity. It can more reliably handle information that requires precise recognition, such as letters and numbers, reducing comprehension errors that occur during voice interactions. In addition, the model supports delegating tasks that require more in-depth reasoning or tool calls to a backend model chosen by the developer. Developers can either use OpenAI's models or integrate third-party models. While the backend model is handling complex tasks, the voice conversation can still continue without interrupting the current interaction. This means that developers can combine real-time voice interaction with the reasoning and tool-calling capabilities of backend models, allowing voice agents to handle complex workflows that require multiple steps while answering user questions. According to OpenAI's official API documentation, the GPT-Live series uses a full-duplex architecture that can directly process input and output audio, reducing the latency caused by the sequential connection of speech recognition (STT), large language models (LLMs), and speech synthesis (TTS) in traditional speech systems. OpenAI has provided technical documentation and access guides related to the GPT-Live-1.1 API. Developers can learn about the model’s capabilities, configuration methods, and specific usage through the OpenAI Developer Platform.
IT之家 11 月 20 日消息,OpenAI 今日宣布向开发者开放 GPT-Live-1.1 API 语音模型,允许第三方开发者将其集成到自己的应用程序中,构建能够实时聆听、回应并执行复杂任务的 AI 语音智能体。 据 OpenAI 介绍,GPT-Live-1.1 是 GPT-Live-1 的升级版本,具备全双工语音交互能力,可以同时聆听和说话,不必等待用户完整说完一句话才开始处理。该模型能够识别用户说话过程中的停顿、打断和语速变化,并实时判断应该继续聆听还是作出回应。 IT之家注:GPT-Live-1.1 自 2026 年 10 月 15 日起已成为 ChatGPT Plus 及更高级别订阅用户的默认语音模型。此次 API 开放,则进一步将相关能力提供给第三方开发者。 在具体能力方面,OpenAI 表示,GPT-Live-1.1 改进了多语言语音处理、指令遵循能力和语音保真度,能够更可靠地处理字母、数字等需要精确识别的信息,减少语音交互过程中出现的理解偏差。 此外,该模型支持将需要更深入推理或工具调用的任务委托给开发者选择的后端模型。开发者既可以使用 OpenAI 的模型,也可以接入第三方模型。后端模型在处理复杂任务时,语音对话仍然可以继续进行,无需中断当前交互。 这意味着,开发者可以将实时语音交互与后端模型的推理、工具调用能力结合起来,让语音智能体在回答用户问题的同时,处理需要多步骤执行的复杂工作流。 根据 OpenAI 官方 API 文档,GPT-Live 系列采用全双工架构,能够直接处理输入和输出音频,减少传统语音系统中语音识别(STT)、大语言模型(LLM)和语音合成(TTS)串联所产生的延迟。 OpenAI 已提供 GPT-Live-1.1 API 相关技术文档及接入指南,开发者可以通过 OpenAI 开发者平台了解模型能力、配置方式和具体使用方法。
想马上用起来?去「AI工具」栏挑一个直接下载,或在「AI教程」里跟着图文步骤做一遍。
去 AI工具 → 看 AI教程 →阅读与点赞数据保存在你的浏览器本地,欢迎留下你的想法。
今日正能量学一点,用一点;今天种下的种子,会长成明天的能力。去免费下载专区 →广告
评论 文明发言,让讨论更有价值