GPQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GPQA (test)
57.2ANLL
4
GPQA
44.5Accuracy
4
GPQA Diamond
84.15Mean@16
4
GPQA full
61.8pass@1
4
GPQA
87.1Extra-task Completion Rate
4
GPQA
27.68EM
4
GPQA Diamond
77.5Score
4
GPQA Diamond
74.33Avg@16
4
GPQA-D v1 (test)
49.9Accuracy
4
GPQA Diamond (test)
47Accuracy
3
GPQA Hard
26.3Accuracy (GPQA Hard)
3
GPQA
10,500.2Average Response Length
3
GPQA-med (Hindi)
43.8Accuracy
3
GPQA Main
37.72Accuracy
3
GPQA Diamond
61.62Accuracy
3
GPQA diamond
47.47Accuracy
3
GPQA Diamond
66.48Average@16
3
GPQA
59.6Accuracy
3
GPQA Diamond
74.49Accuracy
3
GPQA Diamond
60CoT EM
3
GPQA MC All Non-Easy
13.6Recovery Rate
3
GPQA MC (Med. + Hard)
40Recovery Rate
3
GPQA
0.308Log Accuracy
3
GPQA Diamond
37.3Score
3
GPQA
29.91Accuracy
3