PLATFORMS Signal 39
Google launches Gemini 3.5 Transcribe with automatic filler word removal and custom vocabulary support
↗
Google introduced three new Gemini 3.5 audio models including a dedicated transcription model that strips filler words, supports custom vocabularies, and attributes up to three speakers.
For developers building voice interfaces, the public preview in the Gemini API offers transcription with word-level timestamps, speaker attribution, and jargon adaptation that previously required post-processing. The launch also arrives while Google's promised Gemini 3.5 Pro remains unreleased.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
Gemini 3.5 Transcribe succeeds Chirp 3 with improved multilingual performance and lower wording error rates across more than 85 languages.
The models are available in public preview via the Gemini API through AI Studio and Antigravity, with Chrome support pending.
Gemini 3.5 Live and 3.5 Live Experimental add real-time narration and better handling of mid-sentence interruptions for voice chat.
THE CLUSTER