TerraLogic - Tool Selection: leaderboard
Part of TerraLogic (CUHK-Shenzhen and TUM, 2026): 545 scenario-driven geospatial reasoning tasks over optical, SAR and infrared imagery solved through verifiable tool chains, each model driving the HieraPlan reference agent. The score is step-by-step tool-selection accuracy in percent; higher is better.
Metric: Tool-selection accuracy (ToolAcc) (%) in the step-by-step setting over the 545 TerraLogic optical, SAR and infrared remote-sensing tasks, each model driving the HieraPlan tool agent under a ReAct-style protocol in OpenCompass; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 7B Instruct | 74.71 | #2564 |
| 2 | GPT-4o | 72.01 | #741 |
| 3 | Llama 3 70B Instruct | 67.47 | #1676 |
| 4 | DeepSeek V3 | 66.33 | #716 |
| 5 | Qwen 2.5 32B Instruct | 58 | #1215 |
| 6 | InternLM3-8B-Instruct | 57.72 | #2568 |
| 7 | GPT-3.5 | 57.12 | #2417 |
| 8 | Mistral 7B Instruct (v0.2) | 47.05 | #7007 |
| 9 | Gemini 2.5 Flash | 43.31 | #453 |
| 10 | Yi-1.5-6B-Chat | 42.08 | #6399 |
| 11 | Phi-3 Mini 4K Instruct | 41.14 | #6810 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=terralogic-tool-selection · How It Works · Data refreshed daily, snapshot 2026-11-13.