Microsoft AI has unveiled MAI-Transcribe-2-Streaming, its first real-time speech-to-text model, which has quickly risen to the top of its class. Ranking #1 among 38 models on Artificial Analysis, it boasts a word error rate of just 2.5% at 0.13 seconds for final transcripts and 0.12 seconds for first partials.
This innovative model supports continuous language detection across 60 languages, making it a versatile tool for global users. During the introductory period, MAI-Transcribe-2-Streaming is offered at a competitive rate of $0.54 per hour, providing an accessible option for businesses and developers looking to integrate advanced transcription capabilities.
By releasing this model, Microsoft AI aims to enhance real-time communication and accessibility, setting a new benchmark for transcription technology in various applications.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
