A Japanese-optimized large language model developed by NII. Part of the llm-jp-3 series with 8 size variants (150M to 172B). Trained on llm-jp-corpus v3 and fine-tuned using SFT/DPO for instruction following (Instruct2/Instruct3).
Parameters
Undisclosed
Context Window
License
Proprietary
Release Date
2025-03-12
API Pricing
API pricing for this model is not yet available
Strengths
- •Highly optimized for Japanese language understanding and generation.
- •Available in multiple sizes, including a compact 8x1.8B mixture-of-experts model.
- •Versions 13B and below are licensed under Apache 2.0.
Weaknesses
- •Performance on non-Japanese tasks may be limited compared to multilingual models.
- •The largest 172B model requires significant computational resources.
- •Smaller models may have reduced capabilities for complex reasoning.
Use Cases
- •Japanese language chatbots and virtual assistants.
- •Content generation and summarization in Japanese.
- •Research on Japanese language modeling and instruction tuning.