The Verge reports that Google has updated Gemini Audio and introduced the Gemini 3.5 Live, Gemini 3.5 Live Experimental, and Gemini 3.5 Transcribe models. The new transcription capabilities recognize specialized terminology and over 85 languages.
According to the publication, the models are designed for more accurate operation of Google's voice features amidst background noise and in situations where user speech is interrupted. The Verge's headline also indicates the automatic removal of filler words such as 'ums' and 'ahs'.
The practical significance of the update lies in reducing manual processing of transcripts and improving the convenience of voice input, especially in professional environments. However, the available material is presented only as metadata and contains no tests, quality examples, or details about the launch.
editorial commentary
Why it matters
The likely consequence is more convenient application of voice functions in work scenarios where terms and clean transcription are important. The next observable signal will be information regarding availability, languages, and independent quality tests. Significant uncertainty remains due to the single source and metadata lacking detailed results.