Unverified50% confidenceFactExact time
ARC-AGI is a general intelligence benchmark designed by Kaggle founder François Chollet, with human-level performance typically above 85%.
1
Sources
50%
Confidence
Long-term
Relevance
9/8/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedARC-AGI is a benchmark designed by AI safety researcher François Chollet specifically to measure fluid intelligence.79% similarVerifiedFrançois Chollet emphasizes abstract reasoning and generalization ability, and proposed the ARC benchmark as a measure of these capabilities.64% similarUnverifiedNVIDIA's AVO system achieved a perfect score of 100% on the ARC-AGI-3 interactive reasoning benchmark.61% similarUnverifiedARC-AGI contains 400 training problems and 400 test problems, with each problem consisting of 2–3 input-output examples and one test case to be solved.60% similarUnverifiedAgent evaluation benchmark suites include GAIA, AgentBench, and SWE-bench.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/877303API
curl https://kongchang.com/api/v1/knowledge/claims/877303MCP
get_claim(id=877303)