Google launches Gemini 3.5 Transcribe
Analysis based on 6 articles · First reported Aug 26, 2026 · Last updated Aug 28, 2026
The launch strengthens Alphabet Inc.'s competitive position in the AI speech recognition market, potentially increasing adoption of its cloud AI services and driving revenue growth. It may pressure competitors like OpenAI and Amazon to accelerate their own speech-to-text offerings, impacting the broader AI and cloud computing sectors.
On August 26, 2026, Alphabet Inc. announced the launch of Gemini 3.5 Transcribe, a new speech-to-text model designed to convert spoken audio into formatted text while handling filler words, self-corrections, and specialized terminology. The model is available through two APIs: the Live API for real-time streaming with sub-second latency, and the Interactions API for pre-recorded audio with speaker attribution and word-level timestamps. It supports over 85 languages, custom vocabulary, and can identify up to three speakers. According to Artificial Analysis, the model achieves a word error rate of 4.0% for streaming and 2.6% for non-streaming, with a 70% improvement in time to final transcription compared to its predecessor, Chirp 3. The model is integrated across Alphabet Inc. products including Gboard Rambler on Android, the Google Gemini on MacOS, Google Antigravity, and Google AI Studio, with Chrome support coming soon. It is available in public preview for developers and enterprises through the Google Gemini and Gemini Enterprise. The launch follows the broader Gemini 3.5 family release in May 2026 and complements earlier models like Gemini 3.5 Live Translate.
Set up alerts, explore entity relationships, search across thousands of events, and build custom intelligence feeds.
Open Dashboard