xAI's new voice model takes the top spot on Artificial Analysis' speech-to-speech index at the same $0.08 per minute as Voice 2.0.
xAI released Grok Voice 3.0 on October 14, 2026. In independent testing by Artificial Analysis, it ranks first on the firm's speech-to-speech index — the benchmark for models that listen and reply in speech directly, rather than converting speech to text and back.
Pricing didn't change: Grok Voice 3.0 costs $0.08 per minute ($4.80 per hour), the same as Grok Voice 2.0. If you already build on the voice API, the upgrade comes without a budget change.
What this means if you build with Grok: speech-to-speech models matter for anything conversational — phone agents, voice assistants and real-time support. Because the price held steady, the practical question for existing users is quality rather than cost. Test the new model against your current prompts before switching production traffic, especially for anything that handles customer calls.