chinese-llm-benchmark
ReLE (Really Reliable Live Evaluation for LLM, formerly CLiB) is a continuously updated Chinese-language evaluation of large models. According to the README it covers 337 models, from commercial ones (ChatGPT, GPT-5.2, o4-mini, Gemini 3 Pro, Claude 4.5, ERNIE, Qwen3-Max, Baichuan, iFlytek Spark, SenseChat) to open-source ones (Kimi K2, DeepSeek V3.2, MiniMax M2.1, GLM-4.7, Llama 4, Gemma 3, Mistral).
It evaluates seven domains and about 300 sub-dimensions: education, medical and mental health, finance, law and civil service, reasoning and mathematics, language and instruction following, and agent and tool calling. Leaderboards include multimodal, overall (with separate reasoning, commercial and open-source boards), education down to individual school subjects and exams, medical licences, finance qualifications, law and civil-service exams and reasoning sub-tasks. Besides the rankings it publishes a defect library of more than two million model failures for the community to analyse, and offers free evaluation of private models. The README is in Chinese and the repository has no declared licence.
Features
Why choose it
Trade-offs
Maintenance check
- Last commit
- 2026-10-21
- Commits, last 90 days
- 29
- Releases, last 90 days
- 0
- Latest release
- v5.10 (2026-04-21)
- Open issues
- 19
- Recent issues with at least one comment
- 17/30
Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.
Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.
Most discussed open issues
- #14 评测数据太少了吧，这能说明问题？2 comments · 2023-09-14
- #66 有计划增加文生图模型的对比吗？虽然感觉比较主观1 comments · 2025-05-05
- #61 【请求】增加软件开发、编程能力的综合评测。1 comments · 2025-02-26
- #86 要不把你们的微信客服开了吧，根本就不回复消息，什么服务态度？？0 comments · 2026-09-21
- #84 请求添加新模型：Qwen3.7-Flash0 comments · 2026-08-10
Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.
Compatibility
Quick start
Use cases
Alternatives
Comments
No comments yet. Be the first!