新闻
Intelligent transcription with Gemini 3.5 Transcribe
Google DeepMind · 发布于 · 阅读约3分钟
30秒读懂
- 发生了什么
- Google released Gemini 3.5 Transcribe, a speech-to-text model handling real-time streaming and pre-recorded audio with 4.0% and 2.6% word error rates respectively.
- 为何重要
- Developers building voice agents, real-time captioning, or call analytics should evaluate this via Gemini API. Enterprises can access it through Gemini Enterprise Agent Platform.
- 注意
- Multi-speaker support limited to three speakers experimentally. Custom vocabulary and language switching capabilities require testing in production environments.
收听本摘要
- speech
- transcription
- gemini
这条新闻背后的模式
每个模式都讲清楚技术如何运作、何时值得投入，以及在哪里会失效。
The Agent Architect
每周一个模式、一个权衡、一个生产事故案例。为构建智能体系统的人准备的每周简报。
每周一封邮件,一键退订。您的地址仅用于发送简报。