A Japanese-optimized large language model developed by NII. It is one of eight model sizes ranging from 150M to 172B parameters. Trained on the llm-jp-corpus v3 dataset, with SFT and DPO applied via Instruct2/Instruct3. Models under 13B are released under Apache 2.0 license.
Parameters
Undisclosed
Context Window
License
Proprietary
Release Date
2025-05-23
API Pricing
API pricing for this model is not yet available
Strengths
- •Leverages a mixture-of-experts architecture for enhanced capability.
- •Optimized for Japanese through specialized training data.
- •Instruction-tuned for improved task adherence and dialogue.
Weaknesses
- •The mixture-of-experts design can be complex to deploy and optimize.
- •Large model size requires significant computational resources.
- •May inherit biases present in the training corpus.
Use Cases
- •High-performance Japanese language understanding and generation.
- •Complex question answering, summarization, and translation tasks.
- •Research and development in advanced Japanese NLP.