Expired85% confidenceFactTime unknown
Under high-difficulty synthetic adversarial attack scenarios, GPT-5.5's refusal rate for biosecurity-related content dropped by approximately half.
1
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedGPT-5.6的网络安全评测是在降低防护的条件下运行的，评测分数反映模型能力上限而非日常可用水平78% similarUnverifiedGPT-5.5 Instant reduced hallucination rates in the medical and legal domains by approximately half.77% similarUnverifiedGPT-5.6在SWE-bench评测中以不到竞品一半的成本实现更优效果和更高解决率73% similarUnverifiedGPT-5.5 achieved a 46% reduction in error rate compared to GPT-5.4 under Databricks' Agent Harness setup.73% similarUnverifiedOpenAI tested GPT 5.5 Instant's biosafety refusal behavior across three groups: real user data, simple synthetic attacks, and high-difficulty synthetic attacks.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/3407API
curl https://kongchang.com/api/v1/knowledge/claims/3407MCP
get_claim(id=3407)