News
Gemini 3.5 Transcribe Claims 2.6% Word Error Rate and Ships Two Separate APIs for Live and Recorded Audio
Google launched Gemini 3.5 Transcribe on August 26, supporting more than 85 languages with accents and dialects, word-level timestamps, diarization for up to three speakers (more is experimental), custom vocabulary and jargon recognition, and automatic removal of filler words and self-corrections. Google cites Artificial Analysis measurements of 4% average word error rate for streaming and 2.6% for non-streaming, with time to final transcription 70% faster than Chirp 3. Builders get two entry points: a Live API for sub-second bidirectional streaming and an Interactions API for recorded meetings and call logs.
Source
↳ Follow the thread