Model profile
Røst v3 Whisper 1.5B
Transcribes Danish speech. Trained on CoRal v3 conversational and read-aloud speech across dialects.
- Developer / provider
- Alexandra Instituttet / CoRal
- Model type
- Danish speech recognition (ASR)
- Access to weights
- Restricted weights
- Development & adaptation
- Danish-developed / Danish adaptation
What could it be used for?
Test recording transcription or subtitles, checking names and specialist terminology.
These are suggestions for testing, not measured municipal outcomes or a recommendation. Test quality, data flows and operating needs in your own context.
- Output
- Transcribed text
- Access & operation
- Downloadable weights; can run locally with Transformers.
Licence and limitations
Custom CoRal licence based on OpenRAIL-M, with use restrictions. Not unrestricted open source.
Worth checking
The licence restricts impersonation and identity inference, among other uses. Test noise, overlap and local dialects; transcription is not a finished meeting summary.
What is the evidence of quality?
The developer’s CoRal v3 test reports 11.6% character error rate on conversations. This is CER, not WER or a production guarantee.
Version and dates
- Model ID
CoRal-project/roest-v3-whisper-1.5b- Released
- Not separately documented
- Sources checked
Exact launch date not established. Model repository last updated on 19 June 2026; this is not a release date.