GIS agentic spatial analysis on GeoAgentBench (GABench)
89TAO (Reasoning)Claude 3.5
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Claude 3.5Framework=Plan-and-React2026.04 | 89 | 82.95 | 84.94 | 73.02 | 45.5 | 46.11 | 79 | 79.5 | 73.3 | |
| DeepSeek-V3Framework=Plan-and-React2026.04 | 83.79 | 82.2 | 80.64 | 64.48 | 51.16 | 47.34 | 68.5 | 89.5 | 86.2 | |
| GPT-4oFramework=Plan-and-React2026.04 | 79.4 | 83.7 | 79.55 | 61.38 | 38.58 | 40.69 | 59.1 | 81.4 | 79.3 | |
| Gemini-1.5Framework=Plan-and-React2026.04 | 75.99 | 77.08 | 75.11 | 58.23 | 30.52 | 43.19 | 52.9 | 79.6 | 77.5 | |
| GPT-4o-miniFramework=Plan-and-React2026.04 | 74.08 | 71 | 70.95 | 57.8 | 41.1 | 36.15 | 43.4 | 82.6 | 72.2 | |
| Llama-3.1-8BFramework=Plan-and-React2026.04 | 44.26 | 59.34 | 47.11 | 34.07 | 18.93 | 21.59 | 10.1 | 68.2 | 42.4 | |
| Qwen2.5-7BFramework=Plan-and-React2026.04 | 41.86 | 54.03 | 44.79 | 26.49 | 16.14 | 16.9 | 20.1 | 86.9 | 62.8 |