Unverified50% confidenceFactExact time
Perception Bench uses a reverse design approach, extracting questions that all top models answered incorrectly from more than 40 existing benchmarks.
1
Sources
50%
Confidence
Long-term
Relevance
7/22/2026
First Seen
Sources
Related Claims
Unverified在Terminal Bench上，某个模型在模型本身不动、只改框架的情况下，排名从三十名开外冲进前五63% similarUnverifiedWhen the context exceeds 40% of the model's context window, the model's reasoning ability noticeably degrades, exhibiting a decline in intelligence.61% similarUnverifiedOn the Terminal Bench benchmark, a model whose underlying weights were unchanged but whose framework was improved rose in ranking from outside the top thirty to the top five.60% similarUnverified在工具数量超过20个的场景下，工具描述的语义歧义度与模型选择准确率之间存在显著的负相关关系59% similarUnverified少样本方式在格式化输出、分类任务和规则遵循方面的幻觉率可降低40%-70%58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/590288API
curl https://kongchang.com/api/v1/knowledge/claims/590288MCP
get_claim(id=590288)