Expired50% confidenceEventExact time
At the request of the U.S. government, OpenAI for the first time labeled the entire GPT-5.6 model family (from Luna to Sol) as both cybersecurity and biochemical dual high-risk categories.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/8/2026
First Seen
Valid until: 10/6/2026(expired)
Sources
GPT-5.6作弊率史上最高：METR报告揭示AI安全危机
bilibili照观AI6/27/2026
Related Claims
UnverifiedOpenAI's choice to publicly release GPT-5.5 indicates that its internal safety evaluation determined the model's cybersecurity capabilities had not yet exceeded the threshold of unacceptable risk.77% similarVerifiedBefore GPT-5.6 was released, it was first submitted to the U.S. government for review and made available only within a small-scale internal test; this government-first review release process is unprecedented in OpenAI's history and in the entire AI industry.73% similarUnverifiedOpenAI officially and publicly acknowledged that the U.S. government requested restrictions on the release of the GPT-5.6 model.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/230990API
curl https://kongchang.com/api/v1/knowledge/claims/230990MCP
get_claim(id=230990)