xAI launches Grok Voice 3.0, a speech-to-speech model for voice agents, at $0.08 a minute

xAI/Grok

xAI released Grok Voice 3.0 on October 14, 2026, a model that takes spoken input and answers in speech directly, aimed at customer-support and other voice agents. xAI says it sets a new record for resolving support calls and costs less than other frontier voice models.

• Priced at $0.08 per minute ($4.80 an hour), the same as Grok Voice 2.0, and available now in the Voice Agent API.
• Agents set to grok-voice-latest were moved to the new model automatically, with no changes needed on the developer side.
• xAI reports 86.9 on Artificial Analysis's Speech-to-Speech Index, ahead of GPT-Live-1 at 83.2 and Gemini 3.8 Live at 82.6.
• On the tau-Voice support-call test it scores 71.6%; xAI says its latency is 40% lower than GPT-Live-1 at comparable accuracy.
• The benchmark figures are vendor-reported, mostly through Artificial Analysis, and xAI says the model already runs in Tesla vehicles and Grok Bot voice mode.