xAI launches Grok Voice 3.0 for faster, cheaper voice agents
The speech-to-speech model leads Artificial Analysis's index, responds in under a second and keeps the previous version's API price.
xAI has released Grok Voice 3.0, a speech-to-speech model designed for customer support, in-car assistants and other real-time voice agents. It is available through the Voice Agent API and automatically replaces the previous model for developers using the grok-voice-latest alias.
The company says the model responds in under one second and costs $0.08 per minute, unchanged from Grok Voice 2.0. In independent testing by Artificial Analysis, Grok Voice 3.0 scored 86.9 on its Speech-to-Speech Index and 97% on Full Duplex Bench, which measures handling of pauses, interruptions and turn-taking. Artificial Analysis also measured an average cost of $0.025 per task on Big Bench Audio, below the compared OpenAI voice models.
xAI says Grok Voice already handles millions of daily conversations in Tesla vehicles. Starlink call centers use it for about 20,000 calls a day, with roughly 70% resolved without a human, according to the company.
The release matters because it turns frontier voice performance into a deployable API at a competitive price. The benchmark results and production figures are promising, though most usage claims come from xAI and its affiliated companies.
Tesla
xAI
