Unverified50% confidenceOpinionExact time
The 56% score on ARC-AGI-3 still falls considerably short of top human performance and a perfect score.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/4/2026
First Seen
Valid until: 10/18/2026
Sources
Related Entities
Related Claims
UnverifiedOn ARC-AGI-3, Sol Max reached close to 8%, while Opus and Gemini were both below 1%.72% similarUnverifiedOver the past 30 days, the top score on ARC-AGI-3 on the Kaggle platform climbed from 7% to 56%.68% similarUnverified在完全相同的模型下，仅靠Harness策略优化，ARC-AGI-3得分从13.3%提升到38.3%，同时输出Token节省约六倍66% similarUnverifiedPreviously, top models such as GPT-4 had scores on ARC-AGI that hovered in the range of single digits to 30 for a long time.64% similarUnverifiedOpus 5在ARC-AGI 3上的得分是次优模型的三倍62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/973936API
curl https://kongchang.com/api/v1/knowledge/claims/973936MCP
get_claim(id=973936)