@misc{indiciae6d74b23e3610,
  title = {ACTIVE-o3: Empowering MLLMs with Active Perception via Pure Reinforcement Learning},
  author = {Muzhi Zhu and Hao Zhong and Canyu Zhao and Zongze Du and Mingyu Liu and Zheng Huang and Anzhou Li and Hao Chen and Cheng Zou and Jingdong Chen and Ming Yang and Chunhua Shen},
  year = {2026},
  url = {https://arxiv.org/abs/2505.21457},
  note = {Source identifier: 2505.21457}
}