An account number dictated over the phone with noise in the background. An email address said in a hurry. That is exactly where transcription is hardest, and exactly where a mistake costs the most.
SpaceXAI measures right at that spot, and says so.
Fact
The facts: on 18 September 2026 SpaceXAI released Grok Voice Transcribe 2.0. Per the company, the model is twice as accurate as version 1.0 at the same price and ranks first for accuracy among 32 streaming transcription models on the public Artificial Analysis leaderboard. It is built on the audio model behind Grok Voice, which in their words handles tens of thousands of customer support calls a day and the Grok assistant in Tesla vehicles. It transcribes dozens of languages, detects the language itself and follows switches within one recording. Their internal tests include phone calls at 8 kHz, conversations with Grok, dictated codes and emails, and short commands in 19 languages; on short phrases the word error rate falls from 20.6 per cent to 6.8 per cent. The price is $0.10 per hour of audio for batch and $0.20 for streaming, with speaker diarization and timestamps. Atlassian uses it for transcription in Loom. Version 2.0 will soon become the default in the API, and 1.0 will be deprecated in the coming weeks.
Price is the strong part. Streaming transcription is about half of what they show for ElevenLabs and Deepgram in their own chart. For a call centre with thousands of hours a month, that is already a budget line, not a rounding error.
Transcription is judged by names and numbers, not by smooth text.
The ranking, though, is their citation of someone else's leaderboard, and the language chart cites no outside source. And one thing that matters to us: Bulgarian is not in the language chart. That does not mean it cannot transcribe it. It means they do not measure it in front of us.
For conversations in Bulgarian, the only honest test is your own recording with your own noise. Run one hour and count the mistakes in names and numbers.
Official primary sources
SpaceXAI News - Introducing Grok Voice Transcribe 2.0, 18.09.2026The visual is generated code-art, no third-party images.