Silent model
grok-voice-transcribe-1.0 reached end of life on October 2, 2026; requests to the slug are routed to grok-voice-transcribe-2.0 at the same price
What changes
xAI's release notes say grok-voice-transcribe-1.0 is deprecated and reaches end of life on October 2, 2026, and that all requests to that slug are routed to grok-voice-transcribe-2.0 at the same price, with higher accuracy. The source does not say requests fail, so calls keep working but are served by a different model, and transcripts may differ. grok-voice-transcribe-2.0 was announced in the September release notes and is the default model; the Speech to Text docs now list only grok-voice-transcribe-2.0.
Silent: No error. Behavior, results, or cost change quietly.
What to do
Set the model to grok-voice-transcribe-2.0 explicitly (or omit model to get the default, grok-voice-transcribe-2.0) and re-check any tests or downstream logic that depend on exact transcript output.