Reproduce “Can LLM Agents Respond to Disasters? Benchmarking Heterogeneous Geospatial Reasoning in Emergency Operations” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Junjue Wang, Weihao Xuan, Heli Qi, Pengyu Dai, Kunyi Liu, Hongruixuan Chen, Zhuo Zheng, Junshi Xia and 2 more
DORA introduces a benchmark for evaluating agents that combine geospatial tools across disaster-response tasks. It contains 515 expert-authored tasks from 45 real-world disasters, spanning five operational dimensions, with 108 tools and 3,500 expert-verified tool-call steps. Gemini-3.0-Flash scores 53.74% on average final answers versus 80.48% for the gold trajectory, a benchmark gap that does not establish field performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.