88Z AI, 15 October — SpaceXSI has released Grok Voice 3.0, claiming it is the world's most intelligent speech-to-speech model with latency under one second. The model ranks first on the Artificial Analysis comprehensive leaderboard, achieving a 97% score in full-duplex benchmarks and 71.6% in τ-Voice. Its performance is comparable to GPT-Live-1 but with 40% lower latency. Currently, the model supports millions of daily conversations in Tesla vehicles and handles 20,000 calls per day at Starlink customer service cent
At a glance
- Grok Voice 3.0 ranks first on Artificial Analysis with 40% lower latency than GPT-Live-1
- The model handles 20,000 daily Starlink support calls, with 70% requiring no human intervention
The update
SpaceXSI has officially launched Grok Voice 3.0, an end-to-end speech-to-speech model described as the world's most intelligent in its class, with response latency kept under one second. The model tops the comprehensive rankings by Artificial Analysis. Specific data shows it achieves a 97% level in full-duplex benchmarks and a τ-Voice score of 71.6%. While its overall performance is on par with GPT-Live-1, Grok Voice 3.0 successfully reduces latency by 40%. In terms of practical application, the model has been deployed at scale. It supports millions of voice interactions daily within Tesla vehicles and serves Starlink customer service centers, handling 20,000 calls per day. Notably, 70% of these calls are fully automated by the model without human agent intervention, demonstrating high levels of automation capability.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- IT之家Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.