Grok Voice Tops Voice Agent Benchmarks at 67.3%, Elon Confirms Starlink Deployment

xAI's voice model nearly doubles Gemini's score on τ-voice Bench and is already running in production at Starlink — a rare case of a benchmark leader with a real-world deployment to point to.

xAI's Grok Voice has taken the top spot on the τ-voice Bench, a real-time voice reasoning benchmark, scoring 67.3% — nearly 24 points ahead of Google's Gemini at 43.8% and almost doubling OpenAI's GPT Realtime at 35.3%. As @XFreeze reported, the margin isn't close. What makes this more than a leaderboard curiosity is that Elon Musk confirmed in a reply that Grok Voice is already deployed in production at Starlink, as @elonmusk stated plainly: "Grok Voice is used by Starlink right now."

Unlock the full briefing

Get every story in today's briefing, the full archive, and the daily AI intelligence brief.

All stories today

Full archive

Daily brief

Cancel anytime. Payments powered by Stripe.