Changes confirmed medium confidence

xAI Introduces Grok Voice Think Fast 2.0 Ahead of an August 5 Default Switch

The speech-to-speech model is priced at $0.08 per audio minute, with existing latest-alias integrations scheduled to move automatically.

xAI announced Grok Voice Think Fast 2.0 on July 29 as its next-generation speech-to-speech model. The company says the release improves intelligence, transcription accuracy and conversational behavior, while keeping a single audio-minute price for developers.

What changed

Think Fast 2.0 processes voice input and produces voice output. xAI says the model reasons while it is speaking and that this parallel process does not add latency. The announcement also says the new model uses reasoning tokens more efficiently than its predecessor, though this run did not independently verify that claim.

The company published comparative benchmark tables for speech quality, reasoning, conversational dynamics and agentic performance. Those figures are not reproduced here because the linked independent benchmark provider was not fetched before this run reached its evidence-page cap. The official page therefore establishes what xAI announced, but this article does not treat the vendor's comparative ranking as independently checked.

Migration and pricing

The operational deadline is August 5. On that date, xAI says the `grok-voice-latest` alias will move from `grok-voice-think-fast-1.0` to `grok-voice-think-fast-2.0`. Integrations that follow the latest alias should receive the new model without a code change. Developers who need the prior behavior must pin `grok-voice-think-fast-1.0` before the switch.

The listed price is $0.08 per minute of audio. Teams evaluating the model should confirm account-specific access before changing production traffic.

Why it matters

For integrations that follow `grok-voice-latest`, the practical consequence of the announced alias switch is a model change without a code change. Pinning the earlier model gives teams a documented way to preserve existing behavior while they test Think Fast 2.0. The August 5 schedule makes that evaluation time-sensitive for current API users.

Status

Confirmed. xAI's official announcement establishes the model, price and scheduled alias migration. Internal confidence is medium because capability and efficiency claims remain vendor-reported and were not independently tested in this run.

Sources

Update note: Last reviewed 2026-08-01. We will revise this post if xAI changes the August 5 alias migration or published pricing.

Sources

Drafted with AI assistance from source briefs; reviewed for citation completeness and label accuracy.

More Changes coverage