SpaceXAI Launches Grok Voice 3.0, Claims Lowest 1-Second Latency

icon币界网
Share
AI summary iconSummary
SpaceXAI has launched Grok Voice 3.0, claiming it as the fastest voice-to-voice model with under one-second latency. The model scored 86.9 on Artificial Analysis' benchmark, excelling in full-duplex conversations and handling interruptions. It also offers improved cost efficiency and is available via API at the same price as the previous version. With altcoins to watch gaining traction, the Fear & Greed Index remains a key metric for traders monitoring market sentiment.
CoinDesk reports:

SpaceXAI today released Grok Voice 3.0, which the company describes as the world’s most intelligent text-to-speech model. The model has set new benchmarks in handling customer service calls, featuring industry-leading latency of under one second and lower operational costs than other state-of-the-art models. Grok Voice supports millions of conversations daily for Tesla and SpaceX and is now available via API at the same pricing as its predecessor.

In independent tests by Artificial Analysis, Grok Voice 3.0 ranked #1 on the Voice-to-Voice Composite Score with a score of 86.9. This composite evaluates speech reasoning capabilities, τ-Voice metrics, and preference and task completion rates in blind call comparisons. Grok Voice 3.0 achieved a 97% score on the full-duplex benchmark, which measures how naturally the model handles pauses, turn-taking, and interruptions.

τ-Voice evaluates agents' ability to handle various accents, background noise, and interruptions in customer service call scenarios. Grok Voice 3.0 scores 71.6%, performing comparably to GPT-Live-1 (Astra) but with 40% lower latency, establishing a new optimal balance between agent intelligence and response speed.

The official statement also notes that its operational costs are lower than those of comparable models. Artificial Analysis measured the cost of each model completing tasks across major audio benchmarks, finding that Grok Voice 3.0 is less expensive than GPT-Live-1 (Astra, medium version) and GPT-Realtime-2.1.

In the EVA benchmark for ServiceNow’s enterprise voice agent, Grok Voice 3.0 achieved the highest accuracy score of 74% and performed best in the combined evaluation of accuracy and customer experience.

Officially, Grok Voice 3.0 processes millions of conversations daily within Tesla vehicles. Drivers can use Connectors to perform simple commands such as checking their schedule or replying to emails, and leverage Grok Bot to complete complex tasks like ordering coffee or reserving dinner—all without using their hands. Grok Voice 3.0 also supports the voice mode of Grok Bot.

The Starlink Customer Support Center, powered by Grok Voice, handles 20,000 customer calls daily in multiple languages. Approximately 70% of calls are resolved without human intervention, and one-quarter of sales calls result in orders placed before the customer hangs up. A single agent built on Grok Voice 3.0 can perform customer identity verification, hardware troubleshooting, issue refunds, and arrange device replacements.

Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.