Unverified50% confidenceFactExact time
On ARC-AGI-1, before 2024, the accuracy of state-of-the-art AI systems hovered around 30% for a long time, far below the human level of over 85%.
1
Sources
50%
Confidence
Long-term
Relevance
8/11/2026
First Seen
Sources
Prime Agent：能自我改进底层框架的开源编程智能体
producthuntProductHunt Daily8/10/2026
Related Claims
UnverifiedThe average human score on the ARC-AGI test is approximately 85%, while top AI models have long hovered in the 30%–60% range.79% similarUnverifiedThe AI pass rate on SWE-bench Verified was below 40% one year ago and had exceeded 72% as of mid-2025.75% similarUnverifiedHistorical data indicates that inference costs for emerging AI capabilities typically decline by 50%–80% within 18–24 months.71% similarUnverified斯坦福大学2023年的研究表明，主流AI检测工具误判率可高达20%以上68% similarUnverifiedMultiple academic studies in 2023 showed that after typical social media processing, the accuracy of top AI detection tools may drop from over 95% in laboratory settings to 60%–70%.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/732490API
curl https://kongchang.com/api/v1/knowledge/claims/732490MCP
get_claim(id=732490)