xAI releases Grok Voice 3.0, with voice API priced per minute 0.08 USD
BIBIBI
AT A GLANCE
xAI releases Grok Voice 3.0 voice API, reportedly priced at 0.08 USD per minute, with performance and application data disclosed.
Article
xAI releases Grok Voice 3.0, with voice API priced per minute 0.08 USD
PANews October 15 message, xAI releases the voice-to-voice model Grok Voice 3.0，API is now available, priced per minute 0.08 USD, on par with its predecessor. The company officially stated that its response latency is below 1 seconds, and that it ranked first with 86.9 points in the Artificial Analysis voice-to-voice index. The model has been used for millions of daily conversations at Tesla and SpaceX; Starlink customer service handles approximately 20000 calls per day, of which approximately 70%are resolved without human intervention.
PANews states that xAI released the voice-to-voice model Grok Voice 3.0。
02
API which is now available, priced at 0.08 USD per minute, on par with its predecessor.
03
The company officially stated that the model's response latency is below 1 seconds.
04
The company officially stated that the model ranked first with 86.9 points in the Artificial Analysis voice-to-voice index.
05
The report states that the model has been used for millions of daily conversations at Tesla and SpaceX.
06
The report states that Starlink customer service handles approximately 20000 calls per day, of which approximately 70%are resolved without human intervention.
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
This is a model that allows users to communicate directly with the system by voice. Developers can integrate it via API; fees are calculated based on minutes of use. The report also provides latency, evaluation ranking, and actual usage scale data, but several of these figures come from official claims or media accounts.
Why it matters to readers
API Availability and clear pricing mean that developers can evaluate and integrate the voice model; if the low latency, ranking, and large-scale usage data are verified, they can be used to assess its actual service capabilities.
Beginners should focus on the per-minute 0.08 USD usage cost, the official claim of latency below 1 seconds, and whether customer service scenarios can actually reduce human intervention.
Risks and unknowns
The material does not include an official xAI announcement, so the release, pricing, and availability scope cannot be directly verified.
86.9 The claim regarding the score and first-place ranking lacks the original Artificial Analysis evaluation page.
The methodology and specific data underlying the claim of millions of daily conversations by Tesla and SpaceX have not been specified.
Starlink handles approximately 20000 calls per day, and the approximately 70% claim of no human intervention lacks formal disclosure and independent verification.
“Same as the previous generation” does not specify the previous-generation product or the basis for its pricing.
Related Developments
Loading event timeline…
Related concepts
voice-to-voice model
This term is not in the glossary yet. Browse related concepts in the glossary.
