Model profile
DFM Mimir v1
Around one billion parameters, trained from scratch using HRM-Text. Developed by SDU, Aarhus University, University of Copenhagen and the Alexandra Institute.
- Developer / provider
- Danish Foundation Models
- Model type
- Compact Danish and English language model
- Access to weights
- Open weights
- Development & adaptation
- Danish-developed / Danish adaptation
What could it be used for?
Test local text processing, short drafts and bounded Danish-language questions.
These are suggestions for testing, not measured municipal outcomes or a recommendation. Test quality, data flows and operating needs in your own context.
- Output
- Text
- Access & operation
- Downloadable weights; requires the documented tokenizer and attention configuration.
Licence and limitations
Apache 2.0. Open weights; the licence alone does not establish the openness of the entire AI system.
Worth checking
4,096-token context. Not specifically safety-aligned; training languages are Danish and English.
What is the evidence of quality?
Researchers report results on 20 benchmarks. These are not demonstrated municipal production outcomes.
Version and dates
- Model ID
danish-foundation-models/DFM-Mimir- Released
- Sources checked
Official release date.