Noesion Radarnoesion.ai
Release2026-07-29

xAI released Grok Voice Think Fast 2.0, a voice model that listens and speaks directly without converting to text in between. It answers faster than the previous version and makes fewer transcription errors, including on noisy or telephone audio. On 5 August 2026 the generic model name switches to the new version automatically, so anyone wanting to stay on the old one has to name it explicitly first.

What changed

Grok Voice Think Fast 1.0 was the default voice model, slower to first audio and less accurate at transcription.

What it unlocks

Building phone and voice agents that answer faster and transcribe more reliably in noisy or telephone-quality audio.

  • $0.08 per minute of audio
  • 0.70s vs 1.25s time to first audio
  • 82.9% vs 75.7% quality index
  • 1.4x better transcription accuracy

What you need to act on it

  • an xAI API account
  • pin grok-voice-think-fast-1.0 before 5 August 2026 to stay on the old model

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.