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

General Reasoning on MMLU-Pro (pass@1 accuracy)

70.47pass@1 Accuracy

Critique-GRPO (CoT Critique)

14.757229.221143.68558.1489Jun 3, 2025Jul 19, 2025Sep 4, 2025Oct 21, 2025Dec 7, 2025Jan 23, 2026Mar 11, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2025.06
70.47
2025.06
68.25
2026.03
65.8
2026.03
65.6
2026.03
64.8
2026.03
64.3
2026.02
62.8
2026.02
62.5
2026.03
62.5
2026.03
62.3
2026.02
61.6
2026.02
61.5
2026.03
61.4
2026.03
61.2
2026.03
60.8
2026.02
60.3
2026.02
60.3
2026.03
59.9
2026.02
59.3
2026.03
58.5
2026.02
58
2026.02
56.9
2026.02
56.2
2026.02
55.6
2025.06
55.28
2025.10
54.7
2025.10
54.5
2026.02
54.2
2026.02
53.9
2026.02
53.3
2025.10
53.3
2025.10
53
2025.10
52.9
2026.02
52.4
2026.03
52.37
2026.03
51.9
2026.02
51.6
2026.03
51.56
2026.03
51.43
2026.03
50.59
2025.10
50.5
2026.03
50.44
2025.10
49.2
2026.03
46.7
2025.12
46.5
2025.06
46.24
2025.12
44.9
2025.12
44.5
2025.12
44.1
2026.03
43.2
2025.12
42.7
2025.12
42.7
2025.10
42.7
2025.10
42.5
2026.03
42.26
2026.03
42
2025.10
41.7
2026.03
41.46
2025.10
37.7
2026.03
37.54
2025.12
36.7
2025.10
34.1
2025.10
32.7
2026.03
23.41
2026.03
23.25
2026.03
21.24
2026.03
21.03
2026.03
21.02
2025.10
16.9