Grok Voice Think Fast 2.0

model
SpaceXAI

SpaceXAI's second-generation speech-to-speech model, trained with RL for conversational behavior and tool use while reasoning in parallel with speech. It scores 82.9% on Artificial Analysis's Speech-to-Speech Quality Index, 56.5% on tau-voice, and reaches first audio in 0.70 seconds. Evaluation spans 24 languages; price is $0.08 per audio minute.

Announcement

Model Details

Benchmark Scores

Benchmark Score Mode
AA Speech-to-Speech Quality Index 82.9% —
tau-voice 56.5% —
audiospeechmultimodalreasoningagenticproprietary