Unverified50% confidenceBenchmarkExact time

OpenAI在编码欺骗测试中，SO欺骗率从10.4%降到1.3%，Luna从9.5%降到2.8%

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

9/26/2026

First Seen

Valid until: 12/25/2026

Sources

GPT-6与Claude Opus 5.5同日发布：模型价格战全面升级

Related Entities

GPT-6 SOmodel GPT-6 Lunamodel OpenAIcompany

Related Claims

Verified当一次代码审查超过 400 行代码时，审查者的缺陷发现率会急剧下降63% similar UnverifiedOpenAI于2023年推出AI文本检测器，但因准确率仅约26%（假阳性率约9%）而被迫下线63% similar Unverified据OpenViking官方实测数据，输入Token下降幅度在34%到91%之间63% similar UnverifiedDeepSuite基准测试的误报率为44.9%，漏报率为1.2%61% similar Unverified单纯使用'少写代码'提示词可减少33%代码量，但安全指标下降至95%60% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/951258

API

curl https://kongchang.com/api/v1/knowledge/claims/951258

MCP

get_claim(id=951258)