Grok Voice Transcribe 2.0

SpaceXAI's speech-to-text model, twice as accurate as the previous version at the same price.

Handles batch and streaming audio, with speaker separation, word-level timestamps, multichannel transcription, and key term biasing included
Word error rate on short phrases dropped from 20.6% to 6.8%, and multilingual accuracy improved substantially
Ranks first for accuracy among 32 streaming models on the Artificial Analysis leaderboard
Batch transcription costs $0.10 per hour of audio, streaming $0.20 per hour with diarization, timestamps, and key terms included
Existing Speech-to-Text API integrations get the improvement with no code changes. Grok Voice Transcribe 1.0 is deprecated in the coming weeks as 2.0 becomes the default

Turning messy real-world audio into usable text: support calls, interviews, voice commands, spoken names and codes.

Transcription quality is the bottleneck, particularly on short phrases or non-English audio where the last model broke down.

Teams already running the Speech-to-Text API, who get the improvement without touching their code.

Join the newsletter to get ahead with AI.

I share strategies, new modules, free guides, and tools to get you better at using AI in your work.