🧠 Reasoning · llm-stats.com

ARC-AGI-2

Second-generation abstraction and reasoning tasks measuring fluid intelligence on novel visual patterns.

9
测试模型数
GPT-5.5
最佳模型
85%
最高分
OpenAI
提供商
#模型名称ARC-AGI-2提供商Podium 综合得分
1
GPT-5.5
OpenAI
85%
OpenAI54.0
2
77.1%
Google60.6
3
GPT-5.4
OpenAI
73.3%
OpenAI51.0
4
68.8%
Anthropic61.2
5
58.3%
Anthropic38.1
6
GPT-5.2
OpenAI
52.9%
OpenAI49.2
7
42.5%
Meta47.0
8
33.6%
Google42.4
9
31.1%
Google51.9
基准数据聚合自 llm-stats.com。详见方法论