← Back to live feed1 story
The company's new Sonic-3.6 text-to-speech model moved to general availability following a 1085 Elo debut on the Voice Arena US English Leaderboard. Cartesia graduated the tool from preview to GA within 1 week after the model claimed the top ranking and received positive early user feedback.
The developer rebuilt the model architecture from the ground up starting in January rather than tuning existing paradigms. Work on a subsequent model is already underway.
Sign in to suggest edits