AIME
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
AIME 2026
77.7Mean@16
12
AIME25
76.7Pass@16
12
AIME25
75.8Accuracy
12
AIME 25
71.1Avg Score (AIME 25)
12
AIME24
76.9Average Score
12
AIME26
21.46Avg@16
12
AIME25
20Score
12
AIME 24
81.46Accuracy
12
AIME 25
81.04Accuracy
12
AIME25
0.667Accuracy
12
AIME 24
73.3Accuracy
12
AIME25
76Pass@1 Accuracy
12
AIME24
12Average Unique Outputs Count
12
AIME'24
0Avg Max Probability Divergence
12
AIME 25
7.96Speedup
12
AIME 25
42.5Pass@1
12
AIME
56.5Pass@1
12
AIME
0.838WP-AUC
12
AIME 2025
93Pass@k
12
AIME 2025
1.4Average@10
12
AIME 2024
3.6Average Score (a@10)
12
AIME 24
20Training Steps
12
AIME 2025
28.6Peak Accuracy
12
AIME 2024
40.2Peak Accuracy
12
AIME 25
27.5Peak Accuracy
12