Grok Voice 3.0
Explore Grok Voice 3.0, a full-duplex speech-to-speech model built for live voice agents — fast turn-taking, strong call accuracy, and low per-minute cost.
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
The Strengths Behind Grok Voice 3.0
Ranked first on the independent Speech-to-Speech Index at 86.9, Grok Voice 3.0 pairs sub-second response times with call handling proven across millions of daily conversations.
- Best-in-Class Conversation AccuracyArtificial Analysis places Grok Voice 3.0 first on its Speech-to-Speech Index with 86.9, a composite of speech reasoning, τ-Voice tasks, and blind head-to-head preference.
- Full-Duplex in Under a SecondLatency stays below one second while a 97% Full Duplex Bench result shows the model handling pauses, turn-taking, and interruptions the way people do.
- Tested at Tesla and Starlink VolumeGrok Voice 3.0 already fields millions of daily conversations across Tesla and SpaceX, among them 20,000 Starlink support calls where roughly 70% close without a human.
Putting Grok Voice 3.0 to Work
From API access to multilingual call automation — four practical steps for shipping the model inside your own product.
Capabilities That Matter in Production
The benchmarks production voice agents live or die by: accuracy, latency, robustness, and cost per task.
Leadership on an Independent Index
Grok Voice 3.0 holds first place on Artificial Analysis's Speech-to-Speech Index at 86.9 and sets a fresh record for resolving support conversations.
Turn-Taking That Feels Human
A 97% score on Full Duplex Bench shows how gracefully the model handles pauses, hands the floor back and forth, and absorbs interruptions mid-sentence.
Steady Under Accents and Noise
τ-Voice throws support calls at the agent with varied accents, background noise, and interruptions; Grok Voice 3.0 resolves 71.6% of them.
Smart Without Slowing Down
On τ-Voice it matches GPT-Live-1 (Astra) while finishing 40% faster, widening what an agent can be both intelligent and quick.
A Lower Bill per Task
In Artificial Analysis's Big Bench Audio run, spend per completed task lands under both GPT-Live-1 (Astra, Medium) and GPT-Realtime-2.1.
Accuracy That Holds Up in Enterprise
ServiceNow's EVA-Bench puts Grok Voice 3.0 at the top with 74% accuracy and the strongest combined accuracy and customer-experience score.
Grok Voice 3.0: Common Questions
Quick answers on accuracy, latency, per-minute pricing, and how Grok Voice 3.0 behaves on real customer calls.
So what exactly is Grok Voice 3.0?
A speech-to-speech model announced as the most intelligent of its kind, shipping today through the API at the same rate as the previous generation.
How does its accuracy compare?
It sits first on Artificial Analysis's Speech-to-Speech Index at 86.9 and posts 97% on Full Duplex Bench.
What latency should I expect on live calls?
Under one second end to end — class-leading — and 40% faster than GPT-Live-1 (Astra) on τ-Voice.
What does it cost to run?
Nothing changed on price: the Voice Agent API bills $0.08 per minute ($4.80 per hour), exactly as Grok Voice 2.0 did.
Does it cope with genuine support traffic?
It does — Starlink routes 20,000 calls a day through Grok Voice in several languages, and about seven in ten are closed without a human.
Must I update agents already on grok-voice-latest?
No. The migration already ran in the background, so nothing in your integration needs to change.
Ship Your Voice Agent with Grok Voice 3.0
Take the top-ranked speech-to-speech model for a spin: sub-second replies, call resolution at enterprise scale, and Voice Agent API pricing that stays at $0.08 per minute.