xAI releases Grok Voice 3.0 speech model through its API
The new version retains its predecessor’s API price, the company said; performance comparisons are attributed to external evaluators.
Event / data period: 2026-10-14
Published: · Updated:
xAI announced on October 14 that Grok Voice 3.0 was available through its API. The company describes it as a speech-to-speech model and says applications using the `grok-voice-latest` alias were switched automatically. It listed the Voice Agent API price as US$0.08 per minute.
In the Artificial Analysis evaluation cited by xAI, the model scored 86.9 on the Speech-to-Speech Index and 97% on Full Duplex Bench; ServiceNow’s EVA-Bench result was 74% accuracy. These are attributed benchmark results, not guaranteed performance in general use.
Data
xAI releases Grok Voice 3.0 speech model through its API
The new version retains its predecessor’s API price, the company said; performance comparisons are attributed to external evaluators.
- · Occurred
xAI releases Grok Voice 3.0 speech model through its API
The new version retains its predecessor’s API price, the company said; performance comparisons are attributed to external evaluators.
Methodology
Based on the source announcement; no outcomes beyond those stated by the source are asserted.
Updated:
API price and reference duration
Price for the stated one-minute service duration.
| 0.08 | 1 |
Methodology
Based on the source announcement; no outcomes beyond those stated by the source are asserted.
Updated:
Subjects and places
Subject proposals
- Science and technologySuggested
- Preserved editorial category/topic; broad candidate onlyx.ai ↗
Geographic references
Only source-provided geographic references are shown. Opening an external map is optional. Optional story maps use OpenStreetMap. Opening a map sends your IP address and this site’s address to the map provider; location data remains readable as text without loading the map.
- United States · Place role: Coverage · Place type: Country or area · Location precision: RegionalEvidence
- Explicit preserved editorial country/region; no city or national centroid inferredx.ai ↗