SEED-Bench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
SEED-Bench (test)
76.22Accuracy
7
SEED-Bench
76.2Accuracy
7
SEED-Bench2-Plus
70.8Score
7
SEED-Bench
0.94Spearman Correlation (10% subset)
7
SEED-Bench-2-Plus
73Accuracy
7
SEED-Bench (test)
31.04Procedure Understanding
7
Seed-Bench
0.79Seed Score
6
SEED-Bench-2-Plus
75Score
6
SEED-Bench (test)
58.6Accuracy
6
SEED-Bench
76.8Accuracy
6
SEED-Bench all
14.54Performance Gain Sum
6
SEED-Bench 2
67.3SEED-2 Image Score
6
SEED-Bench (test)
52.38TU (Text Understanding)
5
SEED-Bench 2 Plus
0.724Accuracy
4
SEED-Bench
40.9SEED-S Score
4
SEED-Bench-2 Plus
67.85Score
3
SEED-Bench
99.94Average Accuracy
2
SEED-Bench Image
63.8Score
2