DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data
Researchers released Mimir v1, a 1‑billion‑parameter language model built on the Hierarchical Reasoning Model architecture and trained only on permissible post‑training data. It matches larger frontier models on English tasks and achieves state‑of‑the‑art results for Danish.
- Mimir v1 is a 1B‑parameter LLM trained exclusively on permissible data.
- It achieves state‑of‑the‑art performance for Danish and competes with larger models on English tasks.
- All training resources, including weights and scripts, are released openly for the community.
- The work demonstrates that ethical data sourcing can produce high‑quality language models.
The team behind the Hierarchical Reasoning Model (HRM) announced Mimir v1, a 1‑billion‑parameter language model trained from scratch using a curated mix of 161 datasets that are all permissible for post‑training use. By restricting the training data to ethically sourced material, the researchers aim to lower the barrier for open‑source AI development.
Mimir v1 demonstrates performance comparable to much larger frontier models on standard English benchmarks, and it sets a new state‑of‑the‑art for Danish language tasks. The model outperforms the earlier HRM‑Text 1B baseline and competes with larger proprietary systems despite its modest size.
The release includes the model weights, training scripts, and documentation, encouraging the community to build on a transparent and reproducible foundation. This approach highlights the feasibility of high‑quality LLMs without relying on massive, non‑permissible datasets.
The authors hope that Mimir v1 will inspire further research into responsible data practices and expand the availability of strong language models for under‑represented languages like Danish.
Provides an open, ethically trained LLM that can be fine‑tuned for niche applications.
Offers a cost‑effective, high‑quality model for Danish language products without licensing fees.
Shows market potential for responsible AI startups focused on open‑source models.
Serves as a reproducible research example for studying data ethics and model architecture.
Illustrates that powerful AI can be built without compromising data permissions.
- Hierarchical Reasoning Model (HRM)
- An architecture that structures language understanding in multiple reasoning layers to improve performance.
- permissible post‑training data
- Datasets that are legally and ethically allowed for use after a model's initial training phase.
MARC v1: An Open-Source Multi-Agent Framework for Clinical AI Reasoning and Coordination
Open SourceNVIDIA and Local AI Community Fuel Open Source Models and Intelligent Agents
Open SourceMeta returns to open models with Zuckerberg's plan to out-copy China and sell compute by auction
Open SourceI counted every Bun bug before and after the Rust rewrite
Open SourceMeta is back with Muse Glimmer: local, agentic, multimodal, and open source
Naabik'iyati' Committee approves the establishment of the Navajo Nation Artificial Intelligence Policy Working Group WINDOW ROCK, Ariz. – The Naabik'iyati' Committee approved a position statement and the creation of a Navajo Nation Artificial Intelligence P - facebook.com
The Naabik'iyati' Committee has approved the creation of a Navajo Nation Artificial Intelligence Policy Working Group. This group will develop a position statement on AI policy for the Navajo Nation.
As Duke leans into AI, here are the free tools the University offers - The Duke Chronicle
Duke University is making its AI tools available for free to the public, as part of its efforts to lean into AI research and development.
IBM and OpenAI team up to bring AI deeper into the enterprise - IBM
IBM and OpenAI are collaborating to integrate AI into enterprise operations. This partnership aims to enhance business processes with AI capabilities.
AI Does Not Eliminate The Need For Human Judgment - United Nations University
A recent report by the United Nations University emphasizes that AI systems do not replace human judgment, but rather augment it.
UH Maui College receives $660K to enhance AI, cybersecurity education - University of Hawaii System
UH Maui College has received a $660K grant to enhance AI and cybersecurity education. The funding aims to improve digital skills and workforce readiness.
Artificial intelligence is being used in online home listings - KTVN
Artificial intelligence is being used to enhance online home listings, providing potential buyers with more detailed and accurate information. This technology is changing the way people search for homes online.