Current video models succeed on basic understanding but achieve under 25% success on logically grounded generation and near 0% on interactive generation, exposing gaps in multimodal reasoning.
In: The Thirteenth International Conference on Learning Representations (2025)
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2representative citing papers
citing papers explorer
-
How Far Are Video Models from True Multimodal Reasoning?
Current video models succeed on basic understanding but achieve under 25% success on logically grounded generation and near 0% on interactive generation, exposing gaps in multimodal reasoning.
- HyLaR: Hybrid Latent Reasoning with Decoupled Policy Optimization