Introducing Universal-1
Universal-1 is our most powerful speech recognition model. Trained on over 12.5 million hours of multilingual audio data, Universal-1 achieves best-in-class speech-to-text accuracy across four major languages: English, Spanish, French, and German.
