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

Multi-image Understanding on MIBench

72.42Accuracy

Qwen2.5-VL

25.006437.315749.62561.9343Jan 8, 2026Jan 17, 2026Jan 27, 2026Feb 6, 2026Feb 15, 2026Feb 25, 2026Mar 7, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
72.42
2026.03
71.7
2026.01
71.42
2026.01
71.29
2026.01
71.09
2026.03
71.06
2026.03
70.86
2026.01
70.16
2026.03
69.2
2026.01
68.06
2026.01
67.29
2026.03
65.11
2026.03
63.42
2026.03
62.37
2026.03
59.37
2026.01
56.66
2026.03
54.18
2026.01
52.91
2026.03
49.28
2026.01
46.39
2026.01
45.09
2026.01
26.83