tech
Introducing Grok Voice Think Fast 2.0
Introducing our most capable speech-to-speech voice model.
TL;DR
- Grok Voice Think Fast 2.0 offers enhanced intelligence, transcription accuracy, and conversational capabilities.
- It outperforms its predecessor and competing models in various speech-to-speech benchmarks.
- The model shows significant improvement over dedicated transcription models, especially in noisy environments.
- It reasons through queries while speaking, enabling faster tool calls and snappier responses.
- Training focused on human conversational patterns results in shorter sentences, single questions, and fluid interactions.
- A/B testing on Starlink showed increased sales conversion and support containment rates.
- The model will become the default for `grok-voice-latest` on August 5, 2026.
- Pricing is set at $0.08 per minute of audio.