Event ·Multi-source

OpenAI 发布 GPT-Live，实时音频架构实现边听边说

OpenAI 推出 GPT-Live，这是一个面向实时音频的新架构与全栈方案。它能在说话的同时持续聆听，实现边听边说。为在 ChatGPT 规模下达到自然效果，OpenAI 从客户端到模型重建了语音栈。新架构让音频保持连续流动，深层推理与工具调用不会中断对话。

Current read

OpenAI 搞了个 GPT-Live，能让语音助手边听边说，推理和调用工具时不打断聊天，实时对话更自然。

11 sources81° AI heatLast updated 2026/8/3 22:25:53

Evidence

11 sources

Primary sourceGreg Brockman

2026/8/3 22:25:53View source

Related source 1OpenAI

2026/8/3 20:38:34View source

Related source 2小互

2026/8/5 06:18:08View source

Related source 3The Rundown AI

2026/8/3 10:30:23View source

Related source 4ChatGPT

2026/8/4 18:31:35View source

Related source 5Simon Willison’s Weblog

2026/8/1 22:29:44View source

Related source 6IT之家

2026/8/2 04:34:27View source

Related source 7Gary Marcus

2026/8/2 21:43:23View source

Related source 8pandaily

2026/8/3 07:53:05View source

Related source 9Decoder

2026/8/3 10:49:58View source

Related source 10techcrunch

2026/8/3 16:40:55View 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.

Open the story