Google has introduced Gemini 3.5 Transcribe, an advanced speech-to-text model that supports over 85 languages and offers real-time auto-correction capabilities. According to The Decoder, the new model can remove filler words and correct slips of the tongue instantly, improving transcription accuracy and fluidity.
The Decoder also reported that Gemini 3.5 Transcribe achieves a 4.0 percent word error rate in streaming mode and reduces latency by 70 percent compared to Google's previous model, Chirp 3. This significant improvement in speed and accuracy positions Gemini 3.5 as a strong contender in the speech recognition space.
For Japanese markets, where multilingual communication and fast, accurate transcription are increasingly vital in FX, crypto, and equities sectors, Gemini 3.5 Transcribe could enhance real-time data analysis and decision-making efficiency.