Unverified50% confidenceDecision RuleExact time
The winner determination logic is to first sort by average score from highest to lowest; if average scores are tied, completion rate is compared; if still tied, the one with fewer steps wins.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
Unverified在歧义点的四类行为中，'先搜再问'成功率最高，平均通过率达93.4%，远高于'直接猜'的56.5%，'搜多次再猜'为51.5%73% similarUnverified对同一测试用例执行多次调用统计成功比例得到通过率，比二元的通过/失败判断更能反映模型真实表现68% similarUnverifiedFor agent pairs with large reward gaps, the decision disagreement rate is below 1%.68% similarUnverified研究数据表明，在数学竞赛级别问题上，启用深度推理的模型准确率可从不足30%提升至80%以上67% similarUnverifiedBalanced accuracy is calculated as the arithmetic mean of per-class recall rates and is used to address class imbalance problems.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1003676API
curl https://kongchang.com/api/v1/knowledge/claims/1003676MCP
get_claim(id=1003676)