Back to today’s briefing

Story perspectives

Nvidia Launches Game-Changing Speech Model: Transcribes in Seconds!

5/6/2025

45 9

Summary

Story summary

Nvidia has unveiled its groundbreaking Parakeet-TDT-0.6B-v2, a cutting-edge speech recognition model that can transcribe an hour of audio in just one second. With 600 million parameters, it boasts a remarkable 6.05% Word Error Rate, securing the top spot on the Hugging Face Open ASR Leaderboard. This open-source model is set to revolutionize the way developers and businesses handle audio transcription.

1 of 1

Story summary

Nvidia has unveiled its groundbreaking Parakeet-TDT-0.6B-v2, a cutting-edge speech recognition model that can transcribe an hour of audio in just one second. With 600 million parameters, it boasts a remarkable 6.05% Word Error Rate, securing the top spot on the Hugging Face Open ASR Leaderboard. This open-source model is set to revolutionize the way developers and businesses handle audio transcription.