Unverified50% confidenceBenchmarkExact time
在MATH数据集的Level 5竞赛级难题上，前沿模型准确率仍低于60%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/26/2026
First Seen
Valid until: 11/24/2026
Sources
Related Entities
Related Claims
UnverifiedThe 2023 TruthfulQA benchmark showed that even the strongest models at the time did not exceed 60% accuracy on questions requiring precise factual judgment.76% similarUnverifiedJCE所谓的'可靠'实际只有约60%的准确率，因误差累积连续5步后整体成功率仅剩约7.8%75% similarUnverified当综合评估指数中顶尖模型得分接近天花板时会出现基准饱和问题，60分对59分的差距在统计学上通常处于误差范围之内73% similarUnverifiedThe solve rate of current top AI models on SWE-bench remains far below the level implied by their competition performance.71% similarUnverified研究表明，即使在 GSM8K 或 MATH 数据集上接近满分的模型，面对稍作变形或需要跨领域组合的问题时性能下降可达30-50个百分点69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/802593API
curl https://kongchang.com/api/v1/knowledge/claims/802593MCP
get_claim(id=802593)