Updated

Grok Voice 3.0

xAI released Grok Voice 3.0, a speech-to-speech model available through its Voice Agent API. The API is unchanged, and the grok-voice-latest alias now points to version 3.0.

By No Code MBA · Facts checked

Provider
xAI
Announcement date
Status at last check
Generally available
Model family
Grok Voice
Preceding model
Grok Voice 2.0
Capabilities
Audio

What changed

Grok Voice 3.0 is xAI's new speech-to-speech model, replacing Grok Voice 2.0 behind the same API and price. xAI cites Artificial Analysis results of 86.9 on its Speech-to-Speech Index (up from 81.3) and a jump on the tau-Voice support call benchmark from 56.5% to 71.6%, and reports time to first audio under one second. xAI also reports a cost per task of $0.025 on Big Bench Audio.

Access and availability

Connect to the Voice Agent API over WebSocket at wss://api.x.ai/v1/realtime with model grok-voice-latest, which now points to 3.0, or list grok-voice-3.0 explicitly. Ephemeral tokens are supported for client apps, and agents can be tested in the xAI console.

Official access documentation

Pricing

$0.08 per minute ($4.80 per hour) of speech-to-speech, plus $0.004 per text input

Official pricing

What you can build

Compare it with your current voice agent on representative calls, including interruptions and varied accents, and review task completion, latency, and handoff quality.

Limitations

xAI's own table shows it trailing GPT-Live-1 on tau-Voice (71.6% vs 74.5%). Benchmark and customer results are vendor-reported, and the docs do not list rate or session limits.

Sources and release history