GeoAgentBench (Plan-and-React) - Tool Retrieval F1: leaderboard
Metric: Tools-Any-Order F1 (%) between the tools the agent invokes and the ground-truth tool set, on 53 spatial analysis tasks over 117 atomic GIS tools in an executable sandbox (30 steps per task), agents under the authors' Plan-and-React framework (global plan with step-wise reactive execution); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 84.94 |
| 2 | DeepSeek V3 | 80.64 |
| 3 | GPT-4o | 79.55 |
| 4 | Gemini 2.5 Flash | 75.11 |
| 5 | GPT-4o Mini | 70.95 |
| 6 | Llama 3.1 8B Instruct | 47.11 |
| 7 | Qwen 2.5 7B Instruct | 44.79 |
Interactive version: theaggregate.ai/benchmark?slug=geoagentbench-plan-and-react-tool-retrieval-f1 · How It Works · Data refreshed daily, snapshot 2026-10-07.