工具POST
Grok Voice Transcribe 2.0 ranks 5th on non-streaming accuracy
It scores 2.3% AA-WER on non-streaming transcription, down from 4.0% in version 1.0, at roughly 160x speed and $1.67 per 1,000 minutes.
Artificial Analysis
@ArtificialAnlysArchived 2026.09.20Originally posted 2026.09.19
Artificial Analysis @ArtificialAnlysOriginally posted 2026.09.19
On non-streaming transcription, Grok Voice Transcribe 2.0 scores 2.3% on AA-WER, placing #5 of 59 models and improving on Grok Voice Transcribe 1.0 at 4.0%. It sits behind Alibaba's Fun-Realtime-ASR-preview and StepFun's StepAudio 3 ASR at 1.7%, Microsoft AI's MAI-Transcribe-2 at 2.0%, and ElevenLabs Scribe v2 at 2.2%, and ahead of MAI-Transcribe-1.5 at 2.4%, and Gemini 3.5 Transcribe and MAI-Transcribe-1 at 2.6%. It transcribes at a speed factor of around 160x and costs $1.67 per 1,000 minutes.
x.com/artificialanlys/status/2101065585362677984Archived 2026.09.20 11:18