Universal-1

About

Try AssemblyAI's most capable and highly trained speech recognition model trained on 12.5M hours of multilingual audio data. Universal-1 achieves best-in-class speech-to-text accuracy, reduces word error rate and hallucinations, and improves timestamps.