Cohere launches an open-source voice model specifically for transcription
: tors mar 26, 2026 1:30 pm
Relatively light at just 2 billion parameters, the model is meant for use with consumer-grade GPUs for those who want to self-host it. It currently supports 14 languages.
Source: https://techcrunch.com/2026/03/26/coher ... scription/
Source: https://techcrunch.com/2026/03/26/coher ... scription/