Verified75% confidenceFactExact time
By mid-2024, top models had generally exceeded scores of 85–90% on MMLU, causing the benchmark's discriminative power to drop sharply.
3
Sources
75%
Confidence
Medium-term (~90 days)
Relevance
8/4/2026
First Seen
Valid until: 11/2/2026
Sources
Related Claims
VerifiedSome models have already surpassed 90% of the human average score on MMLU, causing their discriminative power to drop sharply.80% similarUnverified基准污染（Benchmark Contamination）导致HumanEval、MMLU等主流基准区分度下降，部分模型HumanEval得分超过95%、MMLU超过90%但实际能力平庸73% similarUnverifiedScale AI在2024年的分析显示，部分模型在MMLU上的分数膨胀可能高达5-10个百分点，因测试集已泄露到训练数据中73% similarUnverifiedTraditional benchmarks such as MMLU and HumanEval have experienced benchmark saturation due to training data contamination; when GPT-4 was released, its MMLU score was approximately 86%, whereas multiple open-source models have now surpassed 90%70% similarUnverifiedThe score gap among multiple frontier models on MMLU has narrowed to the order of one percentage point, driving the emergence of more challenging benchmarks such as MMLU-Pro and ARC-AGI.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/684158API
curl https://kongchang.com/api/v1/knowledge/claims/684158MCP
get_claim(id=684158)