DORA - Spatial Relational Analysis: leaderboard

DORA evaluates model agents answering disaster-response questions with a 108-tool geospatial library under a ReAct loop. This series reports final-answer accuracy for spatial relational analysis; the five tasks cover 515 expert-authored cases across 45 disasters. Only the thirteen agent runs enter this panel; expert gold trajectories and tool-free VLM runs use different protocols. Higher is better.

Metric: Final-answer accuracy (%; ReAct agent with tools). Source: github.com. Saturation forecast: Around February 2027. 13 models tracked. Updated 2 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3.5 397B A17B53.4#255
2Grok 4.1 Fast53.4#429
3GPT-5.450.8#135
4DeepSeek V3.249.15#417
5Gemini 3 Flash49#174
6Gemma 4 31B49#385
7MiniMax-M2.748.53#533
8Claude Sonnet 4.648.23#173
9MiMo-V2-Pro47.43#329
10Step 3.5 Flash44.91#520
11GPT-OSS-120B37.58#707
12GPT-5.4 Nano33.41#645
13Qwen 3.5 35B A3B13.89#490

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=dora-spatial-relational-analysis · How It Works · Data refreshed daily, snapshot 2026-11-17.