Expired50% confidenceBenchmarkExact time
On visual cognition tasks requiring perception + reasoning + image reconstruction, humans can achieve approximately 80%, while SOTA multimodal models achieve only 30% to 50%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/16/2026
First Seen
Valid until: 10/14/2026(expired)
Sources
Related Claims
UnverifiedBy artificially enhancing attention to images or text at the appropriate moment, accuracy on visual-spatial reasoning tasks can be improved from 21% to 50%.68% similarUnverified90% of beginners encounter problems such as visual drift, perspective confusion, and inconsistent style when generating AI videos, and the root cause is almost always in the prompts.67% similarUnverifiedThe 'closure' principle of Gestalt psychology indicates that the human brain actively fills in missing parts of visual information.67% similarUnverifiedAnthropic、DeepMind 等机构的可解释性研究显示，目前只能通过激活分析、注意力可视化等有限手段部分理解模型内部状态64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/533786API
curl https://kongchang.com/api/v1/knowledge/claims/533786MCP
get_claim(id=533786)