A field guide to China’s homegrown, home-runnable LLMs

What it does Awesome-Chinese-LLM is a curated directory of the Chinese open-source large-language-model ecosystem. It collects base models, vertical fine-tunes, training and inference frameworks, datasets, and tutorials. The maintainers bias toward smaller-parameter models that can be privatized and run without enterprise budgets.

The interesting bit The repo’s real utility is its taxonomy. Instead of a flat link dump, it splits the landscape into decision-friendly categories—medical fine-tunes, legal chatbots, pre-training datasets, deployment engines—and includes a comparison table of major base models showing parameter counts, context windows, and commercial licenses.

Key highlights

• Covers 100+ resources, from ChatGLM and Qwen to DeepSeek and MiniMax.
• Breaks down vertical applications across eight domains, including medical, legal, finance, and agriculture.
• Provides a specs-and-licensing cheat sheet for a dozen popular base models.
• Tracks the full lifecycle: training frameworks, inference engines, evaluation suites, and tutorials.
• Open to community contributions via PRs with standardized metadata.

Caveats

• The entire index is in Chinese, so non-readers will need a translator.
• Freshness depends on manual community contributions; the README explicitly solicits PRs for new entries.

Verdict Worth bookmarking if you are building Chinese-language applications or researching model families outside the Western mainstream. Skip it if you need a turnkey SDK or English-only documentation.