Unverified50% confidenceBenchmarkExact time
Even the best-performing models can only recover 30% of the evaluated evidence and 25% of the limitations descriptions.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified25%的风险降幅数字仍需在方法论严格的复制研究中验证，应被视为值得进一步验证的信号而非最终定论70% similarUnverifiedThe proportion of enterprises that can clearly quantify the ROI of AI investment still remains below 30%.66% similarUnverifiedAIME has only 30 samples, and tiny score differences derived from running a small number of seeds are not statistically significant.66% similarUnverifiedOpenScienceCollaboration对100项心理学实验的重复研究发现仅约36%的结果成功复现66% similarUnverifiedSome research estimates that fewer than 20% of strategies selected through backtesting within quantitative funds remain consistently effective in live trading.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/893426API
curl https://kongchang.com/api/v1/knowledge/claims/893426MCP
get_claim(id=893426)