Unverified50% confidenceReleaseExact time
OpenAI最新发布的语音模型能够一边听一边说，几乎实现零延迟的双向对话
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/20/2026
First Seen
Valid until: 10/18/2026
Sources
Related Claims
UnverifiedOpenAI新版语音助手能够在对话中模拟停顿、迟疑、自我修正、轻微清嗓子和无奈语气等非语言信号82% similarUnverifiedOpenAI 官方称新版 ChatGPT Voice 能一边说话一边聆听，比以往更聪明，并且知道何时插话、何时退让80% similarUnverifiedOpenAI新模型采用端到端语音架构，直接以音频作为输入和输出，跳过中间文字转换环节79% similarUnverified领先水平的实时语音转录延迟约在200-400毫秒，业界主流方案包括OpenAI的Whisper、Google Speech-to-Text和AssemblyAI79% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/568264API
curl https://kongchang.com/api/v1/knowledge/claims/568264MCP
get_claim(id=568264)