Stories about Gemini 3.5 Transcribe

Google AI has released Gemini 3.5 Transcribe, a speech-to-text model that reports an average WER of 2.6% across 85+ languages. The model is split into two endpoints: streaming and batch. The streaming endpoint provides sub-second transcription but drops speaker diarization and word timestamps. The batch endpoint keeps both, at half the cost. Google reports a WER of 4.0% for streaming and 2.6% for non-streaming, with 70% faster finalization than Chirp 3.

Google's Gemini 3.5 Transcribe can recognize 85 languages and correct filler words and slips of the tongue in real-time speech-to-text translation. It achieves a word error rate of 4.0% in streaming mode, a 70% reduction in latency compared to the previous generation product Chirp 3. Additionally, tasks can be handed over to other Gemini models via function calls.

Google has announced the launch of Gemini 3.5 Transcribe, which enables AI-driven speech-to-text functionality. This technology will be applied to more Google products, including Chrome. The launch of Gemini 3.5 Transcribe marks Google's further in-depth exploration in the field of speech recognition.

Google's DeepMind has released Gemini 3.5 Transcribe, further enhancing intelligent transcription capabilities. The launch of Gemini 3.5 Transcribe marks another significant breakthrough for Google in speech recognition technology. The improved technology will bring higher accuracy and efficiency to speech-to-text applications.