Unverified50% confidenceFactExact time
The evaluation set consists of a fixed 5,112 samples, mixing ARC-Challenge, ANLI-R3, and stratified subsets of MMLU-Pro and CMMLU.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/4/2026
First Seen
Valid until: 2/2/2027
Sources
Related Entities
Related Claims
VerifiedStandard AI benchmarks include MMLU (a comprehensive knowledge test covering 57 subjects), GPQA (graduate-level professional question answering), SWE-bench (real-world software engineering tasks), HumanEval and MBPP (programming ability), and MATH and GSM8K (mathematical reasoning)69% similarUnverifiedThe research covers four major benchmarks: ARC, GSM8K, HellaSwag, and MMLU.68% similarVerifiedMMLU, HumanEval, and MT-Bench are standardized benchmark test sets widely recognized by the AI community.64% similarVerifiedMainstream large model evaluation benchmarks include MMLU, HumanEval/SWE-bench, MATH, and GPQA63% similarUnverifiedThe BTZSC benchmark evaluated dozens of zero-shot classifiers across 22 datasets, covering NLI models, embedding models, and rerankers, with the associated paper published at ICLR 2026.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1134092API
curl https://kongchang.com/api/v1/knowledge/claims/1134092MCP
get_claim(id=1134092)