A phone book for China’s large-language-model land rush

What it does

This repository is a curated awesome-list that catalogs Large Language Models created in China. The entire README is essentially one giant table mapping companies and research institutes to their model names, home provinces, categories—general, medical, legal, finance—and whether they offer public access. It tracks both closed commercial APIs and open-weight releases, with occasional notes on underlying architectures like LLaMA or BLOOM.

The interesting bit

The value is in the scope. It captures the full spectrum from Baidu and Alibaba down to niche academic fine-tunes for traffic law and mental health. The author treats China’s generative-AI boom as an inventory-management problem, which is probably the right instinct.

Key highlights

• Covers well over 60 entries, marking public availability with a simple ✔ or ✘ and linking to repos or demos where they exist.
• Breaks down models by vertical—finance, law, medicine, education, agriculture—showing where specialization is actually happening.
• Notes architectural lineage when known, so you can see who built from scratch versus who fine-tuned an existing base.
• Accepts community contributions via Issues, positioning itself as a living document.
• Sister repos track benchmarks and open foundation models, suggesting an attempt to map the whole Chinese LLM supply chain.

Caveats

• The list is purely descriptive, not evaluative: no performance numbers, safety tests, or licensing details beyond occasional architecture notes.
• Depth varies row by row; some entries are dense with links, others are just a name and a checkmark.
• The provided source is truncated, so the full current count and newest additions are unclear.

Verdict

Useful if you need a quick map of the Chinese LLM landscape—competitors, collaborators, and curious observers should bookmark it. Skip it if you want model weights, training code, or head-to-head benchmarks; this is strictly reference material.