Speech to Speech / Voice
grok-voice-think-fast-2.0is available as the recommended flagship Speech to Speech model.grok-voice-think-fast-1.0is documented as the previous-generation model.grok-voice-latestis documented as currently aliasinggrok-voice-think-fast-1.0, and will route togrok-voice-think-fast-2.0starting August 5, 2026.- Model lists, best-practices guidance, and Voice REST API reference updated for the new model ID (including session
modeloptions). reasoning.effortdocs no longer limit support to specific voice model IDs.- Section titles renamed from “Speech to Speech API” to “Speech to Speech” on related docs pages.
- See also the announcement and Speech to Speech docs.
Release notes
- Added a July entry for Grok Voice Think Fast 2.0 availability and the
grok-voice-latestrouting date above.
Pages: none added or removed.