---
id: "2026-10-14-grok-voice-3"
url: "https://postcutoff.com/e/2026-10-14-grok-voice-3/"
as_of: "2026-10-14T22:23:00+02:00"
date: "2026-10-14"
date_precision: day
category: model-release
importance: 3
confidence: high
status: [Confirmed]
sources: 2
editor: human editor
human_review: null
version: null
---

As of: 2026-10-14 22:23 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/e/2026-10-14-grok-voice-3/

# SpaceXSI releases Grok Voice 3.0, topping Artificial Analysis' speech-to-speech index at $0.08/min

On Oct 14, 2026 SpaceXSI (formerly xAI) released Grok Voice 3.0, a speech-to-speech model for its Voice Agent API. It scores 86.9 on Artificial Analysis' Speech-to-Speech Index (Oct 13 data), ahead of OpenAI's GPT-Live-1 (83.2) and Gemini 3.8 Live (82.6), answers in under a second and keeps Grok Voice 2.0's price of $0.08 per minute. The grok-voice-latest alias was switched to 3.0 automatically.

## Key facts
- Artificial Analysis Speech-to-Speech Index: 86.9 (Grok Voice 2.0: 81.3; GPT-Live-1: 83.2; Gemini 3.8 Live: 82.6), as reported by SpaceXSI with Oct 13 data
- Big Bench Audio 98.5%; Speech Agent Arena Elo 1070; task success 94.8%; Full Duplex Bench 97.0%; τ-Voice 71.6% (2.0: 56.5%); ServiceNow EVA-Bench 74% accuracy
- Latency: under one second; on τ-Voice, 40% lower latency than GPT-Live-1 (Astra, Medium)
- Price: $0.08/min ($4.80/hour), unchanged from 2.0; cost per Big Bench Audio task $0.025 vs $0.030 for GPT-Live-1 and $0.055 for GPT-Realtime-2.1
- API: Voice Agent API; agents on grok-voice-latest moved to 3.0 automatically; no pinned versioned model id is given on the announcement page
- Deployment claim: Starlink call centres take about 20,000 calls a day in several languages with Grok Voice; about 70% resolved without a human and about one in four sales calls ends in a purchase

## What happened

Two and a half months after Grok Voice Think Fast 2.0, SpaceXSI shipped version 3.0 of its realtime voice model. All
benchmark numbers come from SpaceXSI's announcement; the Artificial Analysis leaderboard itself was not opened for this entry.

## Why it matters

Realtime voice is the interface for call centres and in-car assistants. A same-price upgrade that claims the top independent
speech-to-speech score raises pressure on OpenAI's GPT-Live and Google's Gemini Live pricing.

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 167 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 106 days after its cutoff
- Gemini 4 Argon (training cutoff March 2026): 197 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 136 days after its cutoff

## Sources

1. [SpaceXSI: Introducing Grok Voice 3.0](https://x.ai/news/grok-voice-3) (x.ai, official)
2. [xAI docs: Voice Agent API](https://docs.x.ai/developers/model-capabilities/audio/voice-agent) (docs.x.ai, docs)

## Changes

- 2026-10-14 (filed): Created

## Related

- 2026-07-29: [xAI releases Grok Voice Think Fast 2.0 speech-to-speech model for voice agents](https://postcutoff.com/e/2026-07-29-grok-voice-think-fast-2/index.md)
- 2026-07-08: [OpenAI launches GPT-Live, full-duplex voice models replacing ChatGPT's Advanced Voice Mode](https://postcutoff.com/e/2026-07-08-openai-gpt-live-chatgpt-voice/index.md)
- Models: [Grok Voice 3.0](https://postcutoff.com/m/grok-voice-3-0/) (`grok-voice-3-0`)