Google宣布推出Gemini 3.5 Transcribe，用于AI驱动的语音转文本

<p>在我们等待（<a href="https://arstechnica.com/ai/2026/08/google-says-gemini-has-reached-1b-users-faster-than-any-other-google-product/">可能徒劳</a>）Gemini 3.5 Pro推出的同时，Google正在3.5系列中发布另一个模型。该公司宣布了Gemini 3.5 Transcribe，这是一款旨在通过编辑掉“嗯”和修正来简化语音输入、输出精炼AI文本的AI模型。该模型已经驱动了Pixel 11上的Gboard“Rambler”功能，但即将出现在整个Google生态系统中。</p> <p>根据Google的说法，Gemini 3.5 Transcribe比其之前的语音转文本引擎Chirp 3更快、更准确。这款新的AI模型从语音到最终转录文本的速度应该快约70%，实时语音错误率已降至5.5%。这只比Chirp 3稍好，Google测量为7.32%。尽管如此，使用语音输入时修复打字错误很麻烦，所以这里的任何改进都是有益的。</p> <p><a href="https://cdn.arstechnica.net/wp-content/uploads/2026/08/gemini-3.5-audio-transcribe_fleu.width-1000.format-webp.png"><img width="1000" height="562" src="https://cdn.arstechnica.net/wp-content/uploads/2026/08/gemini-3.5-audio-transcribe_fleu.width-1000.format-webp.png" class="fullwidth full" alt="" decoding="async" loading="lazy" srcset="https://cdn.arstechnica.net/wp-content/uploads/2026/08/gemini-3.5-audio-transcribe_fleu.width-1000.format-webp.png 1000w, https://cdn.arstechnica.net/wp-content/uploads/2026/08/gemini-3.5-audio-transcribe_fleu.width-1000.format-webp-640x360.png 640w, https://cdn.arstechnica.net/wp-content/uploads/2026/08/gemini-3.5-audio-transcribe_fleu.width-1000.format-webp-768x432.png 768w, https://cdn.arstechnica.net/wp-content/uploads/2026/08/gemini-3.5-audio-transcribe_fleu.width-1000.format-webp-384x216.png 384w, https://cdn.arstechnica.net/wp-content/uploads/2026/08/gemini-3.5-audio-transcribe_fleu.width-1000.format-webp-980x551.png 980w" sizes="auto, (max-width: 1000px) 100vw, 1000px"> 图片来源： Google </a></p><p><a href="https://arstechnica.com/ai/2026/08/google-announces-gemini-3-5-transcribe-for-ai-powered-speech-to-text/">阅读全文</a></p> <p><a href="https://arstechnica.com/ai/2026/08/google-announces-gemini-3-5-transcribe-for-ai-powered-speech-to-text/#comments">评论</a></p>