Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Consistent Video Retrieval on COIN (test)
Loading...
51.64
Accuracy
CAST
12.5984
22.7342
32.87
43.0058
Mar 9, 2026
Accuracy
MnR
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
MnR
CAST
Backbone=VideoPrism-B,...
2026.03
51.64
1.9
CAST
Backbone=InternVideo2-...
2026.03
51.03
1.9
CAST
Backbone=GME-Qwen2-VL-...
2026.03
45.68
2.05
CAST
Backbone=Qwen3-VL-Embe...
2026.03
44.87
2.09
Late Fusion (Learned)
Context Modeling=Learn...
2026.03
44.66
2.11
CAST
Context Modeling=State...
2026.03
40.47
2.16
InternVideo2-1B
Backbone=InternVideo2-...
2026.03
17.99
3.36
Late Fusion (Heuristic)
Context Modeling=Fixed...
2026.03
17.85
3.28
Qwen3-VL-Embedding-2B
Backbone=Qwen3-VL-Embe...
2026.03
17.73
3.5
VideoPrism-B
Backbone=VideoPrism-B,...
2026.03
17.6
3.32
GME-Qwen2-VL-2B
Backbone=GME-Qwen2-VL-...
2026.03
17.17
3.44
Early Fusion
Context Modeling=Featu...
2026.03
15.12
2.6
CLIP Baseline
Context Modeling=Conte...
2026.03
14.1
3.91
Feedback
Search any
task
Search any
task