Rakuten open sources its largest Japanese LLM to accelerate local AI innovation, reduce reliance on global models, and enable enterprises, startups, and researchers.
Rakuten Group has open-sourced its latest large language model, Rakuten AI 3.0, under the Apache 2.0 licence, making it freely available via its Hugging Face repository to accelerate AI development and community collaboration.
Positioned as Japan’s largest high-performance LLM, the model is designed to support companies and professionals building AI applications, signalling a strategic move to democratise access to advanced, locally optimised AI and reduce dependence on global, English-centric models.
Rakuten AI 3.0 is optimised for Japanese language tasks and delivers strong performance across writing, code generation, document analysis, and extraction. Compared to earlier models, it demonstrates significantly higher accuracy and robustness.
The model has been benchmarked across Japanese cultural knowledge, history, graduate-level reasoning, competitive mathematics, and instruction following, with results compared against leading Japanese-focused models.
The development is backed by Japan’s GENIAC initiative, with support from the Ministry of Economy, Trade and Industry (METI) and the New Energy and Industrial Technology Development Organization (NEDO), which provided compute resources for training.
“Rakuten is continuously pushing the boundaries of innovation to develop best-in-class LLMs for R&D and deliver best-in-class AI services to its customers. By making the models open to all, Rakuten aims to contribute to the open-source community and accelerate the development of local AI applications and Japanese language LLMs,” said Ting Cai, Chief AI & Data Officer, Rakuten.
“Rakuten is committed to delivering high-quality, cost-efficient models that empower businesses and users. Rakuten AI 3.0, our largest and most competitive model, is an outstanding combination of data, engineering and innovative architecture at scale. By sharing open models, we aim to accelerate AI development in Japan. We are excited for the opportunity to work with the Ministry of Economy, Trade and Industry to foster a collaborative AI development community that drives progress for all.”
Selected for the third term of GENIAC in July 2025, the model was unveiled in December 2025 and later fine-tuned for release, reinforcing open-source AI as a competitive and collaborative pathway for Japan’s ecosystem.
