Google Launches Gemini 3.5 Transcribe AI for Enhanced Speech-to-Text

Google has launched Gemini 3.5 Transcribe, a new AI model designed to significantly improve speech-to-text transcription with enhanced accuracy, real-time capabilities, and support for over 85 languages. The announcement, detailed on August 26, 2026, positions the model as a major upgrade over its predecessor, Chirp 3, with a focus on handling real-world audio challenges like background noise, speaker disfluencies, and complex vocabulary. The model boasts a Word Error Rate (WER) of 4.0 for live streaming and 2.6 for pre-recorded audio, according to metrics from Artificial Analysis.