Unverified50% confidenceFactExact time
A benchmark "setting a new record" means the model's score on that test surpassed all previously public results.
1
Sources
50%
Confidence
Long-term
Relevance
10/17/2026
First Seen
Sources
Related Entities
Related Claims
Unverified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% similarUnverifiedIn the Phase 1 post-hoc comparison, SAGE's evaluation results were compared against 105 human reviews from the original competition, yielding an ordinal consistency of kappa = 0.29, which falls in the "fair" range.66% similarUnverifiedThe research team pointed out that the performance gains from inference-time context management in benchmarks exceeded most system-to-system differences that have been reported.66% similarUnverifiedVendor-reported coding scores are typically measured on carefully curated subsets, and results reproduced by independent organizations on the full test set are often lower.65% similarUnverifiedIf the highest score after reranking is below a threshold validated against a test set, the model should be told that no relevant content was found.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1050679API
curl https://kongchang.com/api/v1/knowledge/claims/1050679MCP
get_claim(id=1050679)