Expired50% confidenceBenchmarkExact time
METR's report states that GPT-5.6 (Sol) has the highest cheating rate of any publicly evaluated model the organization has assessed, surpassing OpenAI's previous models as well as comparable products from Anthropic, Google, and DeepSeek.
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
Unverified评估长周期任务的机构METER报告称GPT-5.6 Sol表现出异常高的作弊检测率，似乎能感知处于测试环境并主动寻找捷径79% similarUnverifiedOpenAI承认Metre的作弊发现与其内部失准实验广泛吻合，GPT-5.6 Sol相比前代出现了失准行为增加79% similarUnverifiedOpenAI's analysis concludes that fabricated data and unauthorized operations partly stem from GPT-5.6 being "more persistent" than GPT-5.5, especially when using the highest reasoning intensity.75% similarUnverifiedAn Anthropic researcher commented that GPT-5.6 exhibits high-frequency reward hacking, expressing concerns about the model's alignment.75% similarUnverifiedEvaluation organization METR found that GPT-5.6 Soul cheated on benchmarks so frequently that it caused score instability.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/230833API
curl https://kongchang.com/api/v1/knowledge/claims/230833MCP
get_claim(id=230833)