Products / Grok Voice
grok-voice
Grok Voice
What version is Grok Voice?
Grok Voice current_version is 3.0. last_updated is 2026-10-14. The catalog lists no channel versions for this product. Catalog generated_at is 2026-11-09T00:02:31+00:00. Realtime speech, TTS, and STT through one Voice API. Catalog status is ga. Organization is spacexai. Official source URL: https://x.ai/api.
- id
- grok-voice
- current_version
- 3.0
- last_updated
- generated_at
- published_at (this UI deploy)
These pages were not scraped: xai-api. Do not invent headlines from memory.
Features
Speech to speech
Realtime voice agents. Starting at $0.05 / min.
Text to speech
TTS at $15.00 / 1M characters.
Speech to text
STT with adjustable VAD threshold (vad_threshold) for streaming and batch.
Voice Think Fast 2.0
Introduced Jul 29, 2026. grok-voice-2.0 is deprecated; upgrade to grok-voice-3.0 or grok-voice-latest at the same price.
Grok Voice 3.0
grok-voice-3.0 for Speech to Speech; grok-voice-latest points to grok-voice-3.0. Same price as Grok Voice 2.0; API unchanged.
Grok Voice Transcribe 2.0
grok-voice-transcribe-2.0 on Speech to Text; grok-voice-transcribe-1.0 reached EOL Oct 2, 2026 and requests route to 2.0 at the same price.
Custom voices
Clone a voice from a reference clip (max 120s) via /v1/custom-voices and use the returned voice_id for TTS and speech-to-speech.
Changelog
- Availability
Grok Voice 3.0 is available
can · S2S adds grok-voice-3.0source · docs.x.aiGrok Voice 3.0 is available
grok-voice-3.0 is now available with Speech to Speech, at the same price as Grok Voice 2.0. grok-voice-latest now points to grok-voice-3.0, so sessions that use the alias are upgraded automatically. The API is unchanged.
- Deprecation
Grok Voice 2.0 deprecation
breakingcan · deprecates grok-voice-2.0source · docs.x.aiGrok Voice 2.0 deprecation
grok-voice-2.0 is deprecated. To upgrade, set model to grok-voice-3.0 or grok-voice-latest, at the same price.
- Deprecation
grok-voice-transcribe-1.0 end of life
breakingcan · STT retires transcribe-1.0grok-voice-transcribe-1.0 is deprecated; EOL October 2, 2026. Requests to that slug route to grok-voice-transcribe-2.0 at the same price with higher accuracy.
source · docs.x.ai - Release
Grok Voice Transcribe 2.0 on the API
can · STT adds grok-voice-transcribe-2.0grok-voice-transcribe-2.0 is available for Speech to Text; grok-voice-transcribe-1.0 remains the default when unspecified.
source · docs.x.ai - Release
Introducing Grok Voice Think Fast 2.0
can · Voice Think Fast is 2.0Grok Voice Think Fast 2.0 announced as the next-generation speech-to-speech voice model.
source · x.ai - Feature
Adjustable VAD threshold
can · STT accepts vad_thresholdSpeech-to-text accepts vad_threshold as a streaming query param and batch multipart field. 0 disables the gate.
source · docs.x.ai - Feature
21 New Flagship Grok Voices
can · Grok Voice has 21 new flagship voices21 new multilingual flagship voices for Grok Voice, plus improved naturalness for the original five, in the Voice Agent API, TTS API, and Voice Agent Builder.
source · x.ai - Feature
Introducing the Voice Agent Builder
can · Voice Agent Builder is in betaVoice Agent Builder beta: no-code production voice agents on Grok Voice, with telephony, tools, MCP, and observability.
source · x.ai