Google · Transcription
Gemini 3.5 Transcribe
Gemini 3.5 Transcribe is Google’s speech-to-text model for recorded audio: automatic detection of 85+ languages including mixed-language speech, speaker labels for up to 8 speakers, word-level timestamps, custom vocabulary, and a smart mode that removes filler words and formats the text. Files up to 20 MB and one hour of audio (30 minutes with speaker labels or word timestamps).
Evidence and sources
Maker HQ: US (according to Google's own website)
All rankings:
#439 on OpenRouter today
#450 on OpenRouter this week
#465 on OpenRouter this monthSources: Google Generative AI API data docs, LiteLLM data, Models.dev data, AI/ML API dataPrimary data from Google Generative AI API
from LiteLLM: variant
from Models.dev: weights availability
from AI/ML API: primary mode, description
Advantages and trade-offs
Compared with 10 representative transcription models
Advantages
Trade-offs
No open weights
The weights aren't public, so you can't run it yourself.
Input & output
as stated by the model's providersText
Audio
Text
Also known as: gemini/gemini-3.5-transcribe, google/gemini-3.5-transcribe, vertex_ai/gemini-3.5-transcribe
Like this one. Tap or hover any dotted value, label or icon to see where it comes from and how it is measured. open details when tapped or hovered. How this data works