Today, we\u2019re announcing Grok Voice 3.0, the world\u2019s most intelligent speech-to-speech model. It sets a new record for resolving customer support calls, has class-leading latency at under one second, and costs less than other frontier models. Grok Voice powers millions of conversations a day across Tesla and SpaceX, and is available today in the API at the same price as its predecessor.<\/p>\n
In Artificial Analysis\u2019s independent testing, Grok Voice 3.0 ranks #1 on the Speech-to-Speech Index, with a score of 86.9. The Index combines speech reasoning, \u03c4-Voice, and preference and task success from blind head-to-head calls. Grok Voice 3.0 also scores 97% on Full Duplex Bench, a measure of how naturally it handles pauses, turn-taking, and interruptions.<\/p>\n
Speech-to-speech benchmark comparison between Grok Voice 3.0 and competing models, per Artificial Analysis.BenchmarkGrok Voice 3.0 (this release)Grok Voice 2.0GPT-Live-1(Astra, Medium)Gemini 3.8 Live(Ext. Thinking, High)<\/p>\n
Overall<\/p>\n
AA Speech-to-Speech Index<\/p>\n
86.981.383.282.6<\/p>\n
Speech Reasoning<\/p>\n
Big Bench Audio<\/p>\n
98.5%97.2%90.1%97.7%<\/p>\n
Human Preference<\/p>\n
Speech Agent Arena Elo<\/p>\n
107010111048990<\/p>\n
Task Completion<\/p>\n
Task Success Rate<\/p>\n
94.8%94.6%87.4%89.1%<\/p>\n
Conversational Dynamics<\/p>\n
Full Duplex Bench<\/p>\n
97.0%95.1%94.9%91.9%<\/p>\n
\u03c4-Voice measures how well agents handle support calls with diverse accents, background noise, and interruptions. Grok Voice 3.0 scores 71.6%, comparable with GPT-Live-1 (Astra), but at 40% lower latency, setting a new frontier in agent intelligence versus responsiveness.<\/p>\n
Agentic Performance vs. Speed<\/p>\n
\u03c4-Voice<\/p>\n
BETTER \u21970%20%40%60%80%0s0.5s1s1.5s2sAgent Success Rate (%)More accurate \u2192Time to first audio (seconds)Faster \u2192Gemini 3.8 Live (Ext. Thinking, High) 68.6%GPT-Live-1 (Astra, Medium) 74.5%Grok Voice 2.0 56.5%Grok Voice 3.0 71.6%0%20%40%60%80%0s0.5s1s1.5s2sAgent Success Rate (%)More accurate \u2192Time to first audio (seconds)Faster \u2192Grok Voice 2.056.5%Grok Voice 3.071.6%Source: Artificial Analysis<\/a><\/p>\n It’s also cheaper to run than comparable models. Artificial Analysis measures what each model costs to complete a task in its Big Bench Audio run, and Grok Voice 3.0 comes in under both GPT-Live-1 (Astra, Medium) and GPT-Realtime-2.1.<\/p>\n Lower cost per task than comparable models<\/p>\n Average cost per task on Artificial Analysis’s Big Bench Audio run, USD. Lower is better.<\/p>\n $0$0.02$0.04$0.06Cost per task (USD)$0.025Grok Voice 3.0$0.030GPT-Live-1(Astra, Medium)$0.055GPT-Realtime-2.1(High)<\/p>\n On EVA-Bench<\/a>, ServiceNow\u2019s benchmark for enterprise voice agents, Grok Voice 3.0 achieves the highest accuracy score at 74% and is the best-performing model when jointly scored across accuracy and customer experience.<\/p>\n \u201cGrok Voice 3.0 sets a new Pareto frontier on EVA-Bench, ServiceNow\u2019s open-source benchmark scoring voice agents on accuracy and experience across 200+ realistic enterprise scenarios. It pairs top accuracy with a natural, responsive conversation, setting a new standard for what enterprises can expect from AI on the front line of customer service.\u201dNitin Aggarwal, Senior Director of Product Management, Agentic Evals and RAI, ServiceNow<\/p>\n Grok Voice 3.0 handles millions of conversations a day in Tesla vehicles. Drivers can use Connectors for simple requests like checking their calendar or answering an email, and Grok Bot for bigger tasks like ordering a coffee or booking dinner\u2014all completely hands-free (see it on X<\/a>). Grok Voice 3.0 also powers Grok Bot\u2019s voice mode.<\/p>\n Starlink call centers use Grok Voice to answer 20,000 customer calls every day across multiple languages. Roughly 70% are resolved with no human in the loop, and one in four sales calls ends in a purchase before the caller hangs up. A single agent built on Grok Voice 3.0 verifies customers, troubleshoots hardware, issues credits, and ships replacement equipment.<\/p>\n