DORA - Disaster Perception and Assessment: leaderboard
Metric: Final-answer accuracy (%; ReAct agent with tools). Source: github.com. Saturation forecast: Around May 2027. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 54.43 |
| 2 | Gemini 3 Flash | 54.19 |
| 3 | MiMo-V2-Pro | 53.68 |
| 4 | Grok 4.1 Fast | 53.07 |
| 5 | GPT-5.4 | 52.85 |
| 6 | MiniMax-M2.7 | 51.87 |
| 7 | Gemma 4 31B | 51.46 |
| 8 | Qwen 3.5 397B A17B | 51.03 |
| 9 | DeepSeek V3.2 | 49.8 |
| 10 | Step 3.5 Flash | 49.7 |
| 11 | GPT-5.4 Nano | 44.4 |
| 12 | GPT-OSS-120B | 42.7 |
| 13 | Qwen 3.5 35B A3B | 14.15 |
Interactive version: theaggregate.ai/benchmark?slug=dora-disaster-perception-and-assessment · How It Works · Data refreshed daily, snapshot 2026-10-02.