See What You Need: Query-Aware Visual Intelligence through Reasoning-Perception Loops
Open paper
Benchmark
Model
Rank
Results
zero-shot-video-question-answer-on-next-qa
CAVIA (GPT-4)
–
Accuracy: 74.5
zero-shot-video-question-answer-on-next-qa
CAVIA (GPT-4.1)
–
Accuracy: 76.1
Rank counts only results with a code link.