awesome-japanese-llm

Why we included this project

Anyone building or evaluating Japanese-language AI will find this index hard to beat as a starting point. It is a volunteer-maintained directory of publicly released Japanese-focused LLMs, organized by architecture, release year, context length, training data, developer, and license, and it also covers the evaluation benchmarks used to compare them. The tables separate text-generation models from multimodal ones and distinguish scratch-trained releases from fine-tuned ones, so you can narrow down candidates without digging through scattered Hugging Face pages and press releases. The maintainers also flag models with non-commercial or vendor-specific licenses, which makes it easy to rule out options that would not work in a commercial deployment. It is a reference list rather than software you run, but for discovery and comparison it does the job well.