DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data
Current large language model development relies on massive, often non-permissible datasets, creating a high barrier for researchers committed to open-source and ethically sourced data. We introduce Mimir v1, a 1-billion-parameter language model based on the Hierarchical Reasoning Model (HRM) architecture, that is trained from scratch and delivers highly competitive performance for English and sets a new state of the art for Danish using only permissible post-training data. Trained on a mixture of 161 datasets, Mimir v1 outperforms the original HRM-Text 1B and competes with larger frontier mode
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 56%Large Language Models Are Still Getting Stronger, but Researchers Face New Bottlenecks in Data, Evaluation, and Safety | Newswise - Newswise →
- PossiblePossibly related (embedding) · 54%Large Language Models (LLMs): Transforming Enterprise AI Development - nerdbot.com →
- PossiblePossibly related (embedding) · 52%Large Language Models (LLMs): Transforming Enterprise AI Development - nerdbot →
- LinkedLinked via arxiv author · 85%Peter Schneider-Kamp →
“DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data”
- LinkedLinked via arxiv author · 85%Jacob Nielsen →
“DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data”
- LinkedLinked via arxiv author · 85%Gianluca Barmina →
“DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data”
- LinkedLinked via arxiv author · 85%Kenneth Enevoldsen →
“DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data”
- LinkedLinked via arxiv author · 85%Lukas Galke Poech →
“DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data”