Expired85% confidenceFactTime unknown
The average human score on the ARC-AGI test is approximately 85%, while top AI models have long hovered in the 30%–60% range.
5
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026(expired)
Sources
Claude 4.5 vs Gemini 3 Pro: A Comprehensive Coding Showdown
bilibiliClaudeEvangelist
Related Entities
Related Claims
UnverifiedOn 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%.79% similarUnverifiedARC-AGI-2由AI安全研究员François Chollet设计，早期GPT-4在该测试上得分不足10%，人类平均得分约85%78% similarUnverifiedAI intent recognition accuracy generally falls between 85% and 95%.72% similarUnverifiedGAIA题目对人类来说相对简单，人类平均正确率约92%，但对顶级AI模型极具挑战性69% similarUnverifiedICIR是平均IC除以其波动，类似夏普比率逻辑，ICIR高于0.5通常被认为值得进一步研究67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/4050API
curl https://kongchang.com/api/v1/knowledge/claims/4050MCP
get_claim(id=4050)