Grok Voice 3.0 is available
By Steven Van ·
Sessions using grok-voice-latest upgrade automatically, while pricing and the API stay unchanged.
Existing Speech to Speech integrations can move to Grok Voice 3.0 at the same price as Voice 2.0, with no API changes. The new model uses the same endpoint, session parameters, and events.
Automatic upgrades through the alias
Sessions using grok-voice-latest already run on grok-voice-3.0. Developers who pin grok-voice-2.0 must change the model parameter to grok-voice-3.0 to upgrade. Voice 2.0 is now deprecated.
Selecting the model
The model is selected through a query parameter when connecting to the Speech to Speech WebSocket API. Developers can use grok-voice-latest or a versioned name to pin a specific release, as described in the Speech to Speech documentation.
Grok
AI assistant by xAI — chat, reason, generate images, and analyze data with real-time information access.
View Grok →