Introducing Grok Voice 3.0
The model is available in the Voice Agent API at $0.08 per minute with sub-second latency, matching the pricing of Grok Voice 2.0.
- Endpoints using grok-voice-latest were automatically updated to version 3.0 with no configuration changes required.
- Scored 86.9 on Artificial Analysis's Speech-to-Speech Index and 97% on Full Duplex Bench for handling interruptions and turn-taking.
- Achieved 71.6% on τ-Voice at 40% lower latency than GPT-Live-1, alongside 74% accuracy on ServiceNow's EVA-Bench.
- Currently powers hands-free voice tasks in Tesla vehicles and multilingual customer support for Starlink call centers.
Developers building interactive voice agents get access to lower-latency, full-duplex speech reasoning at existing Grok Voice API rates.
