SpaceXAI Releases Grok Voice 3.0 With Sub-Second Latency
SpaceXAI has released Grok Voice 3.0, claiming it is the world's most intelligent speech-to-speech model featuring sub-second latency and top performance on independent benchmarks. It is priced at $0.08 per minute via API and is already handling millions of daily conversations in Tesla vehicles and SpaceX customer service. This release represents a major advancement in conversational AI by combining human-like responsiveness with low operating costs and high production reliability at scale. Its deployment in real-world environments like Starlink customer support demonstrates the practical viability of autonomous voice agents for enterprise use. Grok Voice 3.0 ranked first in Artificial Analysis's speech-to-speech composite index with a score of 86.9 and achieved a 74% precision score on ServiceNow's EVA enterprise benchmark. It operates at a lower cost than comparable frontier models while maintaining the same pricing as the previous generation.
## BACKGROUND
Speech-to-speech models process audio inputs directly and generate audio outputs without converting text as an intermediate step, enabling more natural conversational timing and emotional nuance. Benchmarks like Artificial Analysis and EVA evaluate these models across multiple dimensions including speech reasoning, handling interruptions, and executing multi-step enterprise workflows.