Microsoft has launched a new AI transcription model called MAI-Transcribe-2, which it claims is faster and more accurate than competitors such as OpenAI and Google. The model can process audio up to 10 times faster than OpenAI's GPT-Transcribe and 5 times faster than Google's Gemini 3.5 Transcribe, with a higher accuracy rate. MAI-Transcribe-2 ranks first in the FLEURS multilingual benchmark for 60 languages and is able to distinguish speakers, timestamp each word, and recognize specialized terms. The model can also automatically identify languages and handle conversations that switch languages. Transcription results can be selected in verbatim or clean mode depending on user needs.
