Unverified85% confidenceFactTime unknown
GPT-5.2在ARC-AGI基准测试中登顶
1
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026
Sources
GPT-5.2发布:390倍效率提升背后的真相与隐忧
bilibili小溪_YouTube搬运笔记本
Related Entities
Related Claims
UnverifiedGPT-5.6 Sol在ARC-AGI-2上取得92.5%的成绩83% similarUnverifiedGPT-5.6 Sol在编程测试Terminal Bench上达到业界最优(state of the art)75% similarUnverifiedGPT-5.6的首批检查点已开启内部测试,内部代号为Amber Alpha和Beacon Alpha74% similarUnverified在OpenAI官方公布的基准测试中,GPT-5.6在编程领域Terminal Bench拿下新的最高分,生物学基准测试比GPT-5.5更省Token(成绩为OpenAI自行披露,尚未独立验证)74% similarUnverifiedGPT-5.5在OS World Verified基准测试上略逊于Claude尚未发布的新模型73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/3935API
curl https://kongchang.com/api/v1/knowledge/claims/3935MCP
get_claim(id=3935)