xAI has announced Grok Voice Think Fast 2.0, a next-generation speech-to-speech voice model designed to enhance intelligence, transcription accuracy, and conversational capabilities. The update builds on its predecessor by introducing meaningful gains in speech reasoning, conversational ability, and tool use reliability.
- Outperforms dedicated transcription models like Deepgram Nova 3 and ElevenLabs Scribe v2 with a 1.5–2.0× improvement in accuracy across 24 languages.
- Achieves approximately 10× better performance than competitors in noisy settings, such as those with background noise or telephony compression.
- Reasons through queries while speaking in parallel, allowing for snappier tool calls that often execute before the end of the agent's first sentence.
- Utilizes reinforcement learning to produce shorter sentences and ask one question at a time, resulting in simpler and more fluid user interactions.
- Priced at $0.08 per minute of audio, with grok-voice-latest automatically upgrading to version 2.0 on August 5, 2026.
A/B testing on Starlink demonstrated significant increases in sales conversion rates and support containment rates, indicating improved performance across use cases without requiring prompt edits.