Grok Voice Transcribe 2.0 halves errors and keeps the $0.10/hr price
SpaceXAI's new speech-to-text model tops 32 streaming rivals on Artificial Analysis and costs the same as v1.0: $0.10/hr batch, $0.20/hr streaming.
SpaceXAI says Grok Voice Transcribe 2.0 is twice as accurate as 1.0 at the same price. On the Artificial Analysis streaming leaderboard (AA-WER) it now ranks #1 out of 32 models. The biggest gain is in multilingual audio: on short voice-assistant phrases across 19 languages, word error rate fell from 20.6% to 6.8%. SpaceXAI also says it beats every model it tested on 8 kHz customer-support calls.
Pricing stays at $0.10 per audio hour for batch and $0.20 for streaming, with speaker diarization, word-level timestamps and key-term biasing included at no extra cost. It also handles multichannel audio (up to 8 channels), filler-word removal, formatting of numbers, dates and emails, and smart turn detection for voice agents. It detects the language on its own and follows speakers who switch languages mid-recording.
The model is API-only (grok-voice-transcribe-2.0) with no open weights. It becomes the default in xAI's Speech-to-Text API, and 1.0 will be deprecated in the coming weeks.
Why it matters: accurate speech-to-text is now a commodity. At $0.10/hr, xAI is putting price pressure on Deepgram, AssemblyAI and OpenAI in the voice-agent market.
Sources
Primary: the company, paper or repository
Independent coverage
- SpaceXAI Releases Grok Voice Transcribe 2.0 marktechpost.com
Written by an AI pipeline from the sources above. Methodology · Report an error
Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.