Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Mathematical Reasoning on DeepMind-Mathematics

88.4Accuracy

Qwen3-I(4B)

10.50430.72750.9571.173Jun 18, 2024Oct 5, 2024Jan 23, 2025May 13, 2025Aug 30, 2025Dec 18, 2025Apr 7, 2026
Updated 9d ago

Evaluation Results

MethodLinks
2026.04
88.4
2026.04
87.2
2026.04
84.2
2026.04
82.4
2026.04
80.1
2026.04
79.9
2026.04
78.8
2026.04
76.5
2026.04
75.8
2026.04
71.3
2026.04
71.3
2026.04
70.7
2025.03
69.1
2026.04
67.4
2026.04
66.2
2025.03
65.8
2026.04
65.5
2026.04
65.3
2025.03
64.6
2025.03
62.8
2025.03
62.2
2025.03
61.6
2024.06
61.6
2025.03
60.2
2024.06
60.2
2024.06
58.3
2025.03
55.2
2025.03
52.9
2025.03
52.2
2025.03
48
2025.03
45.9
2025.03
45.1
2025.03
43.4
2025.03
42
2025.03
41.9
2025.03
41.7
2025.03
41.5
2025.03
41
2025.03
39.9
2025.03
39.2
2025.03
38.4
2025.03
38.4
2025.03
38
2025.03
36
2025.03
35.6
2025.03
35.4
2025.03
35
2025.03
31.3
2025.03
30.9
2025.03
29.3
2025.03
29.1
2025.03
28.6
2025.03
28
2025.03
27.4
2025.03
27.2
2025.03
26.5
2025.03
25.9
2025.03
21.6
2025.03
21.4
2025.03
18.1
2025.03
17
2025.03
15.4
2025.03
13.5