##article.return## Omni-AutoThink: Adaptive Multimodal Reasoning via Reinforcement Learning Download Download PDF