{"source":"AIMap","schema":"benchmark/v1","updated":"2026-10-22","benchmark":{"id":"c-eval","name":"C-Eval","category":"chinese","unit":"%","higherIsBetter":true,"status":"stale","sources":["ceval"],"primaryVariant":"five-shot","url":"https://cevalbenchmark.com/static/leaderboard.html","year":2023,"desc":{"en":"Chinese evaluation suite of about 13,900 multiple-choice questions over 52 disciplines and four difficulty levels.","zh-hans":"中文综合评测，约 1.39 万道选择题，52 个学科、四个难度等级。"},"measures":{"en":"Average accuracy. Only the initial-release table from the GitHub README is cited here.","zh-hans":"平均准确率。本站只引用 GitHub README 中首发时的表格。"},"limits":{"en":"The live leaderboard (cevalbenchmark.com) could not be reached from our network on 2026-10-10; the README table is from 2023.","zh-hans":"2026-10-10 从本站网络无法访问实时榜单（cevalbenchmark.com）；README 表格是 2023 年的数据。"},"contamination":"high","contaminationNote":{"en":"Validation answers are public; test labels are hidden.","zh-hans":"验证集答案公开，测试集标签不公开。"},"stats":{"results":20,"models":11,"latest":"2023-05-15","bySourceType":{"official":20}}},"sources":[{"id":"ceval","name":"C-Eval (GitHub README, initial release table)","license":"Code MIT; dataset CC BY-NC-SA 4.0","policy":"limited","attribution":"Huang et al., C-Eval","site":"https://cevalbenchmark.com/static/leaderboard.html"}],"results":[{"model":"gpt-4","modelName":"GPT-4","provider":"OpenAI","rawName":"GPT-4","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":68.7,"unit":"%","rank":1,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"chatgpt","modelName":"ChatGPT","provider":"OpenAI","rawName":"ChatGPT","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":54.4,"unit":"%","rank":2,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"claude-v1-3","modelName":"Claude-v1.3","provider":"Anthropic","rawName":"Claude-v1.3","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":54.2,"unit":"%","rank":3,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"claude-instant-v1-0","modelName":"Claude-instant-v1.0","provider":"Anthropic","rawName":"Claude-instant-v1.0","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":45.9,"unit":"%","rank":4,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"glm-130b","modelName":"GLM-130B","provider":"Zhipu AI (Z.ai)","rawName":"GLM-130B","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":40.3,"unit":"%","rank":5,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"bloomz-mt-176b","modelName":"Bloomz-mt-176B","provider":"BigScience","rawName":"Bloomz-mt-176B","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":39,"unit":"%","rank":6,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"llama-65b","modelName":"LLaMA-65B","provider":"Meta","rawName":"LLaMA-65B","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":38.8,"unit":"%","rank":7,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"chatglm-6b","modelName":"chatglm-6b","provider":"Zhipu AI (Z.ai)","rawName":"ChatGLM-6B","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":34.5,"unit":"%","rank":8,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"chinese-llama-13b","modelName":"Chinese LLaMA-13B","provider":"Community (Chinese-LLaMA)","rawName":"Chinese LLaMA-13B","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":33.3,"unit":"%","rank":9,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"moss","modelName":"MOSS","provider":"Fudan (OpenMOSS)","rawName":"MOSS","variant":"five-shot","setting":"Five-shot average, initial-release table (2023)","score":31.1,"unit":"%","rank":10,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"gpt-4","modelName":"GPT-4","provider":"OpenAI","rawName":"GPT-4","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":66.4,"unit":"%","rank":1,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"chatgpt","modelName":"ChatGPT","provider":"OpenAI","rawName":"ChatGPT","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":51,"unit":"%","rank":2,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"claude-v1-3","modelName":"Claude-v1.3","provider":"Anthropic","rawName":"Claude-v1.3","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":50.5,"unit":"%","rank":3,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"bloomz-mt-176b","modelName":"Bloomz-mt-176B","provider":"BigScience","rawName":"Bloomz-mt-176B","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":44.3,"unit":"%","rank":4,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"glm-130b","modelName":"GLM-130B","provider":"Zhipu AI (Z.ai)","rawName":"GLM-130B","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":44,"unit":"%","rank":5,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"claude-instant-v1-0","modelName":"Claude-instant-v1.0","provider":"Anthropic","rawName":"Claude-instant-v1.0","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":40.6,"unit":"%","rank":6,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"chatglm-6b","modelName":"chatglm-6b","provider":"Zhipu AI (Z.ai)","rawName":"ChatGLM-6B","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":38.9,"unit":"%","rank":7,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"llama-65b","modelName":"LLaMA-65B","provider":"Meta","rawName":"LLaMA-65B","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":34.7,"unit":"%","rank":8,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"moss","modelName":"MOSS","provider":"Fudan (OpenMOSS)","rawName":"MOSS","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":33.1,"unit":"%","rank":9,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"},{"model":"chinese-alpaca-13b","modelName":"Chinese-Alpaca-13B","provider":"Community (Chinese-LLaMA)","rawName":"Chinese-Alpaca-13B","variant":"zero-shot","setting":"Zero-shot average, initial-release table (2023)","score":30.9,"unit":"%","rank":10,"rankComputedBySite":false,"date":"2023-05-15","sourceType":"official","source":"ceval","sourceUrl":"https://github.com/hkust-nlp/ceval#leaderboard","collected":"2026-10-10"}]}