Continual Instruction Tuning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
94.6Image-R Score
30
Jun 2, 2026
58.83AP
24
Jun 2, 2026
66.89AP
24
Jun 2, 2026
80.16AP
24
Jun 2, 2026
46.42AP
24
Jun 2, 2026
81.25RS Score
20
Apr 16, 2026
68.1Average Accuracy (AA)
11
Jun 2, 2026
82.36SciQA
10
May 15, 2026
79.72SciQA
8
May 15, 2026
61.42CoIN Average
7
Jul 3, 2026
66.9Average Accuracy (AA)
6
Jun 2, 2026
83.59ScienceQA Accuracy
6
May 19, 2026
91.07ImageNet-R Score
6
Mar 23, 2026
74.6Average Accuracy
3
Apr 21, 2026