Unverified50% confidenceBenchmarkExact time
MMLU contains more than 14,000 multiple-choice questions across 57 subjects, and current state-of-the-art models have already exceeded a score of 90%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/5/2026
First Seen
Valid until: 12/4/2026
Sources
Related Entities
Related Claims
VerifiedMMLU contains approximately 15,000 multiple-choice questions across 57 subject areas; HumanEval, designed by OpenAI, contains 164 Python programming problems; MATH contains 12,500 competition math problems; and GPQA covers graduate-level science questions.74% similarUnverifiedLFW基准于2007年提出，包含5749个身份13000张图像，顶尖模型已接近饱和超过99.8%59% similarUnverifiedCOCO 数据集包含约12万张训练图像、80个类别、超过150万个目标实例，由微软研究院发布56% similarUnverifiedA 70-billion-parameter model corresponds to approximately 1.4 trillion training tokens under the Chinchilla optimum52% similarUnverifiedThe community ECC solution contains 63 Agents, all in MD file format, of which 6 use Opus for deep reasoning, 55 use Sonnet for everyday tasks, and 1 uses Haiku.51% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/858933API
curl https://kongchang.com/api/v1/knowledge/claims/858933MCP
get_claim(id=858933)