SpaceXSI releases Grok Voice 3.0, claiming top spot on speech-to-speech index
SpaceXSI announced Grok Voice 3.0 on October 14, calling it the world's most intelligent speech-to-speech model, with sub-one-second latency and availability in the Voice Agent API at $0.08 per minute — the same price as Grok Voice 2.0 — with apps on grok-voice-latest migrated automatically. According to Artificial Analysis testing dated October 13, the model ranks first on the Speech-to-Speech Index with 86.9, ahead of Grok Voice 2.0 at 81.3, OpenAI's GPT-Live-1 at 83.2 and Google's Gemini 3.8 Live at 82.6, and the company cites an average cost per task of $0.025, below the comparison models. SpaceXSI also says Grok Voice handles millions of conversations a day in Tesla vehicles, while Starlink call centers use it for about 20,000 customer calls daily in multiple languages, roughly 70% of which are resolved without a human.
Voice agents are a leading deployment path for enterprise customer service, so lower latency and per-task cost could shape vendor choices in that segment.