Unverified80% confidenceFactTime unknown
ARC-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%
1
Sources
80%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
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.78% similarUnverifiedOn the ARC-AGI-3 benchmark, GPT-5.6 Sol scored 8%, while most other models scored below 2%; Anthropic did not run Fable due to prohibitively high costs.73% similarUnverified自2023年GPT-4发布以来，AI在HumanEval等主流编程基准测试上的通过率从不足50%跃升至90%以上71% similarUnverified2023年人类在GAIA测试中的平均成功率约为90%，而当时最强的GPT-4在Level 1上仅勉强达到15%68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/13611API
curl https://kongchang.com/api/v1/knowledge/claims/13611MCP
get_claim(id=13611)