math Math · llm-stats.com

FrontierMath

Research-grade mathematics problems crafted by professional mathematicians.

7
테스트된 모델
GPT-6 Astra
톱 모델
97.6%
최고 점수
OpenAI
제공업체
#모델명FrontierMath제공업체Podium Score
1
97.6%
OpenAI86.2
2
89%
OpenAI80.8
3
84.9%
OpenAI66.4
4
78.6%
OpenAI49.1
5
GPT-5.4
OpenAI
47.6%
OpenAI51.0
6
GPT-5.2
OpenAI
40.3%
OpenAI49.2
7
GPT-5.5
OpenAI
35.4%
OpenAI54.0
벤치마크 데이터는 llm-stats.com에서 집계. 자세한 내용은 방법론.
0 / 4