Grok Voice 3.0
Explore Grok Voice 3.0: a speech-to-speech model built for live calls, with near-instant replies and a voice agent API priced from $0.08 per minute.
Support
Pro AI Tools
Explore elite tools including the Suno AI Music Generator.
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator
Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
Why Grok Voice 3.0 Leads Speech AI
An independent index places Grok Voice 3.0 first among speech-to-speech models, pairing top call-resolution accuracy with latency low enough for real conversation.
- Ranked First for Call AccuracyArtificial Analysis puts it at 86.9 on its Speech-to-Speech Index — a composite of speech reasoning, τ-Voice results, and blind head-to-head preference.
- Replies in Under a SecondLatency stays below one second, and a 97% Full Duplex Bench score shows the model handles pauses, turn-taking, and interruptions the way people do.
- Tested at Starlink and Tesla VolumeIt already runs millions of daily conversations across Tesla and SpaceX, including 20,000 Starlink support calls a day — roughly seven in ten closed without a person.
How to Put Grok Voice 3.0 to Work
From your first API call to multilingual support automation, here is the four-step path to shipping with the model.
Grok Voice 3.0 Benchmarks That Matter
Accuracy, latency, robustness, and cost per task — the numbers that decide whether a voice agent survives production.
First Place on an Independent Index
The model posts a new high for resolving support calls and takes the top spot on Artificial Analysis's Speech-to-Speech Index at 86.9.
Conversation That Flows Naturally
A 97% Full Duplex Bench result captures how well it handles pauses, passes the floor back and forth, and absorbs interruptions mid-sentence.
Holds Up Against Accents and Noise
τ-Voice runs agents through support calls with varied accents, background noise, and interruptions; Grok Voice 3.0 reaches 71.6%.
Smart Without the Lag
On τ-Voice it matches GPT-Live-1 (Astra) while finishing 40% faster, widening what an agent can be both clever and quick.
Cheaper per Completed Task
In Artificial Analysis's Big Bench Audio run, spend per task lands under both GPT-Live-1 (Astra, Medium) and GPT-Realtime-2.1.
Built for Enterprise Support
ServiceNow's EVA-Bench gives it the highest accuracy at 74% and the strongest combined accuracy and customer-experience score.
Grok Voice 3.0 Questions, Answered
Quick answers on accuracy, response speed, per-minute pricing, and how Grok Voice 3.0 performs on live customer calls.
What exactly is Grok Voice 3.0?
It is a speech-to-speech model positioned as the most intelligent of its kind, and it is available in the API now at the same rate as the previous version.
How accurate is it on benchmarks?
It holds first place on Artificial Analysis's Speech-to-Speech Index at 86.9 and scores 97% on Full Duplex Bench.
How quickly does it respond mid-call?
Latency sits under one second, and on τ-Voice it answers 40% faster than GPT-Live-1 (Astra).
What does Grok Voice 3.0 cost to run?
Nothing changed on price: the Voice Agent API charges $0.08 per minute, or $4.80 per hour, the same as Grok Voice 2.0.
Does it work on real customer calls?
It does — Starlink routes 20,000 support calls a day through it in several languages, and about seven in ten are resolved without a human.
Will existing grok-voice-latest agents break?
No. If your agent already targets grok-voice-latest, the upgrade rolled out quietly and nothing in your setup has to change.
Put Grok Voice 3.0 on Your Next Call
Spin up a voice agent on the top-ranked speech-to-speech model — sub-second replies, enterprise call resolution, and $0.08/min API pricing that hasn't moved.