OpenCALM

OpenCALM is a family of openly released Japanese language models from CyberAgent, built on the GPT-NeoX architecture and trained on Japanese Wikipedia, Japanese mC4, and Japanese CC-100. It shipped in sizes from small and medium up through 1B, 3B, and 7B (6.8B) variants, giving Japanese researchers a permissively shared-license foundation at a time when most capable models were English-first.

The models are released under CC BY-SA 4.0, which allows commercial use with attribution and share-alike on adaptations, a notable contrast to the research-only licenses common for early Japanese LLMs. OpenCALM seeded a wave of Japanese fine-tunes, including the University of Tokyo's Stormy. CyberAgent later followed it with CALM2-7B, a Llama-based model on 1.3T tokens under Apache 2.0.