Grok Voice Transcribe 2.0 doubles transcription accuracy at constant cost
September 18, 2026
SpaceXAI released Grok Voice Transcribe 2.0, a speech-to-text model trained on noisy, multilingual audio datasets. It achieves twice the accuracy of version 1.0 while maintaining the same price point, specifically optimized for difficult real-world environments like phone lines and vehicle cabins.
HOW THIS AFFECTS YOU
●
builderYou can improve transcription reliability in noisy environments without increasing your inference budget.
●
founderThis lowers the cost barrier for deploying high-accuracy voice agents in physical products and customer support.