Visual Question Answering
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
0.864COMET Score
4
May 7, 2026
0.168Sentence BLEU-4
4
May 7, 2026
25.9Sentence BLEU-4
4
May 7, 2026
0.131Sentence BLEU-4
4
May 7, 2026
0.37Sentence BLEU-4
4
May 7, 2026
12.7Sentence BLEU-4
4
May 7, 2026
84.2VQA Accuracy
4
Apr 29, 2026
95.9VQA Accuracy
4
Apr 29, 2026
61.3Accuracy (VQA)
4
Apr 29, 2026
63.6Accuracy
4
Apr 29, 2026
26.1Score (Noise)
4
Apr 24, 2026
36.5Accuracy (Noise Corruption)
4
Apr 24, 2026
60.4Accuracy (Noise)
4
Apr 24, 2026
90.5Overall Accuracy
4
Apr 24, 2026
79.03Accuracy
4
Apr 23, 2026
61.5ROUGE-L
4
Apr 23, 2026
0.62Relative Accuracy Drop (%)
4
Apr 21, 2026
68.8Accuracy
4
Apr 21, 2026
41.2Accuracy
4
Apr 21, 2026
45.13Pass@1
4
Apr 21, 2026
54.5Accuracy
4
Apr 15, 2026
61Accuracy
4
Apr 15, 2026
48.7Accuracy
4
Apr 15, 2026
2BLEU-4
4
Mar 19, 2026
6.2Reasoning Score
4
Mar 17, 2026