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

Multi-image Reasoning on Mantis-Eval

78.3Overall Score

InternVL2.5-38B

34.20445.65257.168.548Aug 9, 2024Aug 28, 2024Sep 17, 2024Oct 7, 2024Oct 27, 2024Nov 16, 2024Dec 6, 2024
Updated 1mo ago

Evaluation Results

MethodLinks
2024.12
78.3
2024.12
77.6
2024.12
77
2024.12
75.6
2024.12
73.7
2024.12
71.4
2024.12
69.6
2024.12
69
2024.12
67.7
2024.12
66.8
2024.12
65.4
2024.08
63.1
2024.08
62.7
2024.12
62.7
2024.12
62.7
2024.12
61.3
2024.08
59.5
2024.12
54.8
2024.08
51.2
2024.12
51.2
2024.08
48.9
2024.12
48.4
2024.12
46.1
2024.08
45.6
2024.08
45.2
2024.12
39.6
2024.08
39.2
2024.08
35.9