@misc{indiciae89d4a5639e35,
  title = {Light-Omni: Reflex over Reasoning in Agentic Video Understanding with Long-Term Memory},
  author = {Chang Nie and Jiaju Wei and Junlan Feng and Chaoyou Fu and Caifeng Shan},
  year = {2026},
  url = {https://arxiv.org/abs/2607.05511},
  note = {Source identifier: 2607.05511}
}