Verified85% confidenceFactTime unknown
ARC-AGI is a benchmark designed by AI safety researcher François Chollet specifically to measure fluid intelligence.
5
Sources
85%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Claude Sonnet 4.6全面评测：性能逼近Opus的知识工作AI新标杆
bilibili英文白斑马
Related Entities
Related Claims
UnverifiedARC-AGI is a general intelligence benchmark designed by Kaggle founder François Chollet, with human-level performance typically above 85%.79% similarUnverifiedARC-AGI was proposed in 2019 by former Google engineer François Chollet, whose original design intent was to provide a general reasoning evaluation set unaffected by the distribution of training data.75% similarUnverifiedARC (Abstraction and Reasoning Corpus) was proposed by François Chollet in 2019 and was specifically designed to test fluid intelligence, requiring the induction of rules from very few examples and their application to new instances.72% similarUnverifiedEPRI一个试点项目正在测试让AI查看焊缝检查产生的海量数据，看它能否标记出值得关注的区域，该项目仍保留人在回路中审查AI的所有建议。71% similarVerifiedFrançois Chollet emphasizes abstract reasoning and generalization ability, and proposed the ARC benchmark as a measure of these capabilities.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/7412API
curl https://kongchang.com/api/v1/knowledge/claims/7412MCP
get_claim(id=7412)