Cohere launches an open source voice model specifically for transcription
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.
Topic
Model Specifically
Model Specifically: 1 recent articles from 1 sources, related entities, and follow-up coverage in one page.
Articles
1
Sources
1
Last update
Mar 26, 2026 at 13:30
Ad slot
Topic page ad slot
A reserved partner slot for relevant products, services, and editorial sponsorships.