SpaceXAI released Grok Voice Transcribe 2.0 for batch and streaming speech-to-text. Artificial Analysis reports a 2.3% word error rate on its public index versus 4.0% for version 1.0, while the API price remains $0.10 per audio hour for batch and $0.20 for streaming. The model adds word confidence, diarization, eight-channel transcription, key-term biasing and automatic language switching, and is already used for every Atlassian Loom video.
SpaceXAI halves transcription error without raising API prices
SpaceXAI released Grok Voice Transcribe 2.0 for batch and streaming speech-to-text. Artificial Analysis reports a 2.3% word error rate on its public index versus 4.0% for version 1.0, while the API price remains $0.10 per audio hour for batch and $0.20 for streaming. The model adds word confidence, diarization, eight-channel transcription, key-term biasing and automatic language switching, and is already used for every Atlassian Loom video.
Imported from the existing research archive. Migration preserves the recorded judgment and is not a fresh verification.
What this means for you
For voice-agent and media pipelines, this is a material quality gain at a price low enough to change build-versus-buy calculations. Trial it on noisy domain audio and proper nouns before switching defaults; public aggregate accuracy does not replace workload-specific error, latency and privacy testing.
Evidence & related material
Technical details & scores
| Model Id | grok-voice-transcribe-2.0 |
|---|---|
| Public Accuracy | Artificial Analysis WER 2.3% versus 4.0% for Grok Voice Transcribe 1.0 |
| Pricing | $0.10/audio-hour batch; $0.20/audio-hour streaming |
| Features | word timestamps and confidence, speaker diarization, up to 8 channels, 100 key terms, smart turn detection |
| Multilingual | dozens of languages; detects mid-recording language switches |
| Adoption | Atlassian Loom uses the model for every video transcription |