Grok Voice Transcribe 2.0 announced with batch and streaming support
TestingCatalog reports that the multilingual speech-to-text model ranks first for accuracy among 32 streaming models on the Artificial Analysis leaderboard.
TLDR
SpaceXAI announced Grok Voice Transcribe 2.0, calling it “the world’s most accurate speech transcription model.” TestingCatalog reports that it ranks first for accuracy among 32 streaming models on the Artificial Analysis leaderboard. TestingCatalog also lists batch and streaming support, speaker diarization to distinguish speakers, multichannel transcription, text formatting and filler word removal among its features.
Combined views
14K
1 Source, first seen ago