Grok Voice Transcribe 2.0

2026-09-18 · Publisher : xAI · Domain : Artificial Intelligence

SpaceXAI releases Grok Voice Transcribe 2.0, a transcription model that is twice as accurate, multilingual, with word-level timestamps, speaker separation and multichannel support, at the existing pricing.

What changes

Accuracy announced as "twice" that of version 1.0.
Word-level timestamps, diarization and multichannel audio.
Available via API on September 17; version 1.0 remains the default.

### Specifications - Identifier: grok-voice-transcribe-2.0 (exact name not verified). - Modalities: audio to text, batch and streaming. - Availability: xAI API. ### Pricing | Mode | Price | Note | |---|---|---| | Batch | 0.10 USD/hour | - | | Streaming | 0.20 USD/hour | - | ### Benchmarks No official figures published by the vendor in the sources consulted. ### Key takeaways An upgrade with no price increase for transcription use cases.

Grok Voice Transcribe 2.0

What changes

Links