GeoDisaster - Answer Accuracy: leaderboard
Answer accuracy on GeoDisaster (2026), the share of disaster geo-intelligence questions whose final answer a GPT-5.5 judge marks correct when the model works as a single tool-using agent over Earth-observation and GIS evidence. Percent; higher is better.
Metric: Answer accuracy (%; final answer judged correct) on the GeoDisaster test instances (verified operational disaster geo-intelligence questions over optical and SAR imagery, raster masks, vector geometries, road networks and exposure layers, in five task families), each model as a single agent with structured prompting, the shared 18-tool registry and execution budget; a GPT-5.5 judge scores every trajectory; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 61.99 | #49 |
| 2 | GPT-5 | 60.48 | #167 |
| 3 | GPT-4o | 60.19 | #741 |
| 4 | O4 Mini | 26.64 | #353 |
| 5 | Qwen 2.5 7B Instruct | 2.53 | #2564 |
| 6 | Qwen 2.5 VL 7B Instruct | 1.5 | #1833 |
| 7 | Qwen 3 4B 2507 Instruct | 1.16 | #2163 |
| 8 | Llama 3.1 8B Instruct | 0.25 | #4246 |
| 9 | Mistral 7B Instruct (v0.3) | 0.25 | #6653 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=geodisaster-answer-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-13.