SpaceXAI said on 18 September that Grok Voice Transcribe 2.0 is selectable on the Speech-to-Text API at 0.10 dollars per hour for batch and 0.20 dollars per hour for streaming, matching version 1.0. Version 1.0 stays the default until a switch planned in the coming weeks. DataNorth reports the same price hold and the same default-if-omitted rule, and notes that xAI did not publish a public version-to-version benchmark beside the price.
SpaceXAI announced Grok Voice Transcribe 2.0 on 18 September with batch transcription at 0.10 dollars per hour of audio and streaming at 0.20 dollars per hour, the same published list prices as version 1.0, with diarization, timestamps and key terms included. DataNorth reports the same 0.10 and 0.20 dollar Speech-to-Text list prices.[1], [2]
Existing Speech-to-Text integrations can select the new model without code changes, SpaceXAI says. Version 1.0 remains the default; 2.0 is to become the default and 1.0 deprecated in the coming weeks, and pinning grok-voice-transcribe-1.0 keeps the older model. DataNorth says 2.0 is selectable while 1.0 stays the default if the model is omitted.[1], [2]
SpaceXAI claims 2.0 is twice as accurate as 1.0 on its own evaluations and that Artificial Analysis ranks it first among 32 streaming models; those figures are vendor-presented. DataNorth notes xAI has not published a version-to-version public benchmark with the price hold. This card stays inside the desk news cards' figures and dated stamps; it does not claim a second field census or an independent lab rerun. The reader sees the publisher's own-page development together with the second publisher's headline or lede on the same event.[1], [2]
