SEED-Bench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
SEED-Bench Image
73.7Accuracy
18
SEED-Bench L3 OOD R1
49.3Accuracy
16
SEED-Bench L2 OOD R1
51.6Accuracy
16
SEED-Bench-R1 L1 In-Dist.
50.5Accuracy
16
SEED-Bench (val)
58.8Accuracy
16
SEED-Bench 1
73.5Image Accuracy
15
SEED-Bench Spatial
66.28Accuracy
15
SEED-Bench Image Part
75.9Accuracy
15
SEED-Bench all (val)
65.6Accuracy
14
SEED-Bench-I Image Subset
71Accuracy
13
SEED-Bench IMG 2023a
77Accuracy
13
SEED-Bench Image
100Average Accuracy
11
SEED-Bench 2-Plus
72Accuracy
11
SEED-Bench-2-Plus
53.3Accuracy Score
10
SEED-Bench Image
77.39Accuracy
10
SEED-Bench Image
77.3Accuracy
10
SEED-Bench (cleaned)
88.64Overall Score
10
SEED-Bench
62.8Accuracy
10
SEED-Bench Video
53.2Accuracy
10
Seed-Bench (test)
66.9Overall Score
10
SEED-Bench 2-Plus
70.22Accuracy
9
SEED-Bench video
61.3Accuracy
8
SEED-Bench all
60.52Accuracy
8
SEED-Bench SEED^I Image-only
72.5Accuracy
8
SEED-Bench Image
67.9SEEDI Score
7