# TestingCatalog — SpaceXAI says Grok Voice Transcribe 2 doubles accuracy

- Company: TestingCatalog (testingcatalog.com)
- Announced: 2026-09-19T15:01:23+00:00
- Category: new-model
- Coverage: not counted
- Announcement: yes
- Group: models
- Source: https://www.testingcatalog.com/spacexai-says-grok-voice-transcribe-2-doubles-accuracy/
- Record: https://forck.live/items/12325-spacexai-says-grok-voice-transcribe-2-doubles-accuracy
- Subject: AI models / agents / unreleased features
- Models affected: Grok Voice Transcribe 2.0, grok-voice-transcribe-1.0
- Pricing: $0.10 per hour for batch transcription and $0.20 per hour for streaming, including diarization, timestamps, and key terms

SpaceXAI released Grok Voice Transcribe 2.0, a speech-to-text model it says is twice as accurate as version 1.0 at the same price. The model ranks first among 32 streaming models on the Artificial Analysis accuracy leaderboard, and its multilingual short-phrase word error rate fell from 20.6% to 6.8%. Existing API integrations receive the update without code changes, and version 1.0 is due for deprecation in the coming weeks.

## Evidence

Verbatim from https://www.testingcatalog.com/spacexai-says-grok-voice-transcribe-2-doubles-accuracy/:

> Word error rate on the multilingual short-phrase set fell from 20.6% to 6.8%, while the model led every tested system on telephony audio.

---

Record: https://forck.live/items/12325-spacexai-says-grok-voice-transcribe-2-doubles-accuracy
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md