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

Science Question Answering on ScienceQA IMG (test)

98.4Accuracy

InternVL2.5-8B

30.07247.81165.5583.289Nov 21, 2023Apr 13, 2024Sep 5, 2024Jan 27, 2025Jun 21, 2025Nov 12, 2025Apr 6, 2026
Updated 11d ago

Evaluation Results

MethodLinks
98.4
97.1
96.7
2026.04
96.6
95.4
2026.04
94.3
2026.04
93.8
2026.04
91.2
2026.04
90.5
2026.04
90.1
89.2
88.9
2026.04
88.2
85.7
2024.02
82.1
2026.04
81.2
81
2024.02
80.7
2024.02
80.1
2024.02
79.9
2024.02
78.3
2024.02
78.3
2024.02
73.6
2024.02
73
2024.02
72.6
2023.11
71.6
2024.03
71.6
2024.02
71.2
2024.02
71.2
2024.02
70.9
2024.02
70.6
2024.02
70.6
2024.02
70.4
2024.02
70.1
2024.02
70
2024.02
70
2024.02
69.4
2024.02
69
2024.02
68.7
2023.11
68.4
2024.03
68.4
2024.02
68.4
2026.04
68.4
2023.11
68.2
2024.03
68.2
2023.11
67.1
2024.03
67.1
2023.11
66.8
2024.03
66.8
2026.04
66.7
2024.03
65.4
2024.02
64.5
2023.11
63.1
2024.03
63.1
2024.02
63.1
2026.04
63.1
2024.03
61.2
2023.11
61
2024.03
61
2024.02
61
2023.11
60.5
2024.02
60.5
2026.04
58.8
2026.04
57.3
2026.04
56.7
2026.04
54.7
2026.04
53.5
2026.04
51.1
2024.03
49.2
2026.04
46.8
2026.04
44.8
2026.04
39.6
2023.11
38.5
2026.04
32.7