Today, xAI announced the launch of Grok Voice Think Fast 1.0, a cutting-edge voice model designed for intricate, multi-step workflows, achieving the highest score on the Tau Voice Bench. This advanced model excels in real-world environments, managing challenges such as noise and accents, and is already utilized in Starlink’s customer support and sales, resolving 70% of inquiries autonomously across various tools and languages. Grok Voice Think Fast 1.0 features innovative architecture that allows for real-time reasoning without sacrificing response speed, making it an ideal solution for high-stakes scenarios that require precise data entry and seamless interaction.

xAI: xAI is an AI company developing advanced models and APIs for enterprise applications, including voice agents that handle complex workflows. It announced Grok Voice Think Fast 1.0, a flagship voice model topping the Tau Voice Bench leaderboard for real-world conditions like noise and accents. The model powers deployments such as Starlink’s customer support and sales via a dedicated phone line.
Starlink: Starlink is a satellite internet service provider enabling global connectivity through its constellation. It collaborates with xAI to deploy Grok Voice for phone-based sales and customer support at +1 (888) GO STARLINK, managing multilingual interactions and high-stakes decisions like hardware replacements.
Grok Voice: Grok Voice is xAI’s voice agent platform designed for multi-step workflows in customer support, sales, and data entry with low latency and high accuracy. The new Grok Voice Think Fast 1.0 version excels in real-time reasoning, tool orchestration, and handling messy audio inputs across 25+ languages. It leads benchmarks and supports enterprise use cases like appointment booking and troubleshooting.

Benchmark Leadership: Grok Voice Think Fast 1.0 tops the Tau Voice Bench leaderboard for full-duplex voice agents under realistic conditions including noise, accents, and interruptions.
Real-World Deployment: The model powers Starlink’s autonomous customer support and sales, resolving most inquiries without human intervention across numerous tools and languages.
Architecture Innovation: It uses parallel processing to separate acoustic interaction from background reasoning, enabling subsecond responses without latency compromises.